Illnesses resembling toxic shock syndrome (TSS) have been reported since 1927, but TSS was first defined as a disease in 1978 by Dr. James K. Todd and colleagues using criteria designed for epidemiologic studies ( Box 64.1 ). TSS is characterized by fever, hypotension due to massive loss of fluid from capillaries into the interstitial space, and subsequent multisystem end-organ dysfunction. It has unique clinical manifestations not generally noted in septic shock, including diffuse erythroderma, delayed desquamation of the palms and soles, and conjunctival and pharyngeal hyperemia.
Clinical Findings
- •
Fever: Temperature of ≥38.9°C (102°F)
- •
Rash: Diffuse macular erythroderma
- •
Desquamation: 1 to 2 weeks after onset of illness, particularly on palms, soles, fingers, and toes
- •
Hypotension: Systolic blood pressure ≤90 mm Hg for adults; <5th percentile by age for children <16 years; orthostatic drop in diastolic blood pressure of ≥15 mm Hg from lying to sitting; orthostatic syncope or orthostatic dizziness
- •
Involvement of three or more of the following organ systems:
- •
Gastrointestinal: Vomiting or diarrhea at onset of illness
- •
Muscular: Severe myalgia or creatinine phosphokinase level greater than twice the upper limit of normal for the laboratory
- •
Mucous membrane: Vaginal, oropharyngeal, or conjunctival hyperemia
- •
Renal: Blood urea nitrogen or serum creatinine greater than twice the upper limit of normal or 5 or more white blood cells per high-power field in the absence of a urinary tract infection
- •
Hepatic: Total bilirubin, aspartate transaminase, or alanine transaminase greater than twice the upper limit of normal for the laboratory
- •
Hematologic: Platelets <100,000/mm 2
- •
Central nervous system: Disorientation or alterations in consciousness without focal neurologic signs when fever and hypotension are absent
- •
- •
Negative results on the following tests, if obtained:
- •
Blood, throat, or cerebrospinal fluid cultures; blood culture may be positive for Staphylococcus aureus
- •
Serologic tests for Rocky Mountain spotted fever, leptospirosis, or measles
- •
Case Classification
- •
Probable: A case with 5 of the 6 clinical findings above
- •
Confirmed: A case with all 6 of the clinical findings above, including desquamation, unless the patient dies before desquamation could occur
When first described, TSS was associated with menses, tampon use, and a distinctive type of Staphylococcus aureus . Patients with menses-associated TSS (menstrual TSS) had cervicovaginal colonization with strains of S. aureus that made a toxin initially called staphylococcal enterotoxin F or pyrogenic exotoxin C but now known as TSST-1 . Toxic shock–like syndromes were subsequently recognized in cases not associated with tampons (nonmenstrual TSS) and in patients infected with toxin producing Streptococcus pyogenes . TSST-1, staphylococcal enterotoxins, and streptococcal pyrogenic exotoxins A and B are now recognized to be “superantigens.” These superantigens are potent stimuli of T-cell proliferation, causing massive release of cytokines (interleukin-1 [IL-1], IL-2, tumor necrosis factor-α [TNF-α], and TNF-β) that mediate the systemic manifestations of the disease.
Epidemiology
Surveillance and Incidence
The first 12 cases of epidemic TSS were identified and reported to the Centers for Disease Control and Prevention (CDC) between July 1979 and January 1980. In May 1980, shortly after surveillance for TSS began the CDC reported findings of the first 55 cases, 95% of which occurred in women and had onset during menstruation. By June 1980, case-control studies statistically linking the occurrence of menstrual TSS with tampons had been completed. Superabsorbent tampons, particularly the Rely brand and those containing polyacrylate, were found to have a high association with TSS. Between 1983 and 1994, the CDC received reports of 2509 confirmed cases; 95% occurred in females, 89% had an onset of TSS associated with menstruation, and 99% occurred in tampon users.
Between 1980 and 1996, the incidence of menstrual TSS was highest in the ages 15 to 25 years (median age, 21 years, 1979–80; 20 years, 1981–86; 25 years, 1987–96). Roughly one-third of cases occurred in adolescents 15 to 19 years of age ( Fig. 64.1 ). A striking 97% of cases occurred in whites, who comprise 83% of the U.S. population. Although TSS cases occurred throughout the United States, the five states of Wisconsin, Minnesota, Colorado, Utah, and California accounted for 44% of the total reported cases but represented only 16% of the U.S. population. In addition to differences in surveillance and menses-related practices, the distribution in age, race, and geography have been attributed to the prevalence of toxin-producing organisms and antibody to exotoxins.
Geographic differences continued even with an increase in the relative proportion of nonmenstrual cases.
The mean age of patients with nonmenstrual TSS (27 years through 1982, 30 years in 1986) is significantly higher than that for menses-associated cases. The number of confirmed cases reported in children younger than 10 years of age is surprisingly low, particularly in light of their low antibody titers to TSST-1 and more frequent nasal colonization with TSST-1–positive strains. The racial distribution is less pronounced, with 87% occurring in whites, who make up 83% of the U.S. population, and the ratio of males to females with nonmenstrual TSS is closer to that in the general population.
Since 1980, when the rate of menstrual TSS was as high as 12.3 per 100,000 women of menstruating age, the incidence has steadily declined to less than 1 case per 100,000 ( Fig. 64.2 ). The menstrual TSS–related mortality rate has decreased from 5.5% in 1979–80 to 1.8% in 1987–96. The striking reduction in the incidence of menstrual TSS has been attributed to changes in composition and usage of tampons and the impact of publicity on early recognition of symptoms. Since 1982, when the U.S. Food and Drug Administration (FDA) required that information on TSS risks appear on tampon packages, tampon composition has changed to cotton and rayon combinations with the removal of additives. Products with high absorbency decreased from 42% to 1%, and tampons containing polyacrylate were withdrawn by major manufacturers.
The incidence of postoperative cases of TSS after all types of surgeries has been estimated to be 3/100,000. For ear, nose, and throat surgery, the incidence is higher (16.5 per 100,000). Overall the average annual incidence of nonmenstrual TSS has been estimated to be 0.32 per 100,000 (95% confidence interval [CI], 0.12–0.67). Largely due to the declining incidence of menstrual TSS, the relative proportion of nonmenstrual cases reported has increased from 14% during 1979–86 to 42% in 1994. The case-fatality ratio for nonmenstrual cases has remained between 5.3% and 8.5%.
Risk Factors for Toxic Shock Syndrome
The primary risk factors necessary for the development of TSS include colonization or acquisition of a toxin-producing strain of S. aureus, absence of protective antitoxin antibody, and an infected site with or without a foreign body. The main distinction between menstrual and nonmenstrual TSS is the nature of the infected site. While menstrual TSS–involves cervicovaginal colonization or infection associated specifically with tampon use, cases of nonmenstrual TSS reported to the CDC include nonsurgical staphylococcal cutaneous and subcutaneous infections, 22%; postsurgical infections, 15%; childbirth or abortion, 15%; vaginal infections occurring at times other than during menses, 6%; vaginal contraceptive sponge use, 5% ; diaphragm use, 6%; and no obvious focus (presumed vaginal or pharyngeal colonization), 31% ( Box 64.2 ).
- I.
Colonization of infection with toxin-producing Staphylococcus aureus
- II.
Absence of protective antitoxin antibody
- III.
Infected site
- A.
Primary S. aureus infection
- •
Carbuncle
- •
Cellulitis
- •
Dental abscess
- •
Empyema
- •
Endocarditis
- •
Folliculitis
- •
Mastitis
- •
Osteomyelitis
- •
Peritonitis
- •
Peritonsillar abscess
- •
Pneumonia
- •
Pyarthrosis
- •
Pyomyositis
- •
Sinusitis
- •
Tracheitis
- •
- B.
Postsurgical wound infection
- •
Abdominal
- •
Breast
- •
Cesarean section
- •
Dermatologic
- •
Ear, nose, and throat
- •
Genitourinary
- •
Neurosurgery
- •
Orthopedic
- •
- C.
Skin or mucous membrane disruption
- •
Burns (chemical, scald, etc.)
- •
Dermatitis
- •
Influenza
- •
Pharyngitis
- •
Postpartum (vaginal delivery)
- •
Superficial/penetrating trauma (insect bite, needle stick)
- •
Viral infection
- •
Varicella
- •
- D.
Foreign body placement
- •
Augmentation mammoplasty
- •
Catheters
- •
Contraceptive sponge
- •
Diaphragm
- •
Surgical prostheses/stents/packing material/sutures
- •
Tampons
- •
- E.
No obvious focus of infection (vaginal or pharyngeal colonization)
- A.
Colonization With Exotoxin-Producing Staphylococcus Aureus
S. aureus is present in 20% to 40% of nasal cultures. In menstruating women, 26% are colonized with S. aureus in at least one of three body sites (nose, vagina, or anus). Vaginal carriage rates vary from 7% in nonmenstruating women to 33% in menstruating women and are higher during menses than at midcycle. In healthy individuals, 14% to 39% of S. aureus isolates produce TSST-1, and 7% to 14% produce staphylococcal enterotoxin B. TSST-1–positive S. aureus is present in 1% to 5% of vaginal cultures in women, 7% of nasal cultures in hospitalized patients, and 18% of nasal cultures in children. Overall up to 7% of healthy individuals at any given time are colonized at a mucosal site with TSST-1–positive S. aureus . TSST-1 producing isolates has been found in 28% of patients bacteremic with S. aureus without TSS. Presumably these patients had circulating antibody to TSST-1 that prevented the development of TSS despite S. aureus infection.
The majority of toxin-producing S. aureus isolates in both menstrual and nonmenstrual TSS cases produce TSST-1. Of the strains producing TSST-1, 60.6% also produce an enterotoxin. A clone producing both TSST-1 and staphylococcal enterotoxin A is associated with 88% of menstrual and up to 54% of nonmenstrual TSS cases. Isolates that produce TSST-1 in conjunction with staphylococcal enterotoxin C have been associated with severe respiratory tract TSS-associated infections. Staphylococcal enterotoxin B is never co-expressed with TSST-1. TSST-1–negative strains that produce staphylococcal enterotoxin B account for 38% of nonmenstrual TSS isolates. Patients with nonmenstrual disease infected with these TSST-1–negative strains may have a higher mortality rate.
While toxin producing staphylococcal strains may cluster within families, living units, and hospital settings, the occurrence of TSS clusters is a rare event. Probable TSS cases have developed within 24 hours in a husband and wife and in two mother-daughter pairs. Nosocomial acquisition and transmission of TSST-1–positive organisms have been described.
Absence of Protective Antibody Levels
Antibody formation to exotoxins is thought to result from mucosal colonization with TSST-1–positive S. aureus strains. Adults and children with persistent nasal carriage of a TSST-1–positive strain have high levels of antibody to TSST-1. Antibody titers are higher in women colonized with toxigenic S. aureus whether the colonization is persistent or transient. Patients in whom TSS develops have significantly lower levels of antibody to staphylococcal exotoxins than the general population. Following the episode of TSS, the antibody response to TSST-1 is typically absent or delayed.
The prevalence of antibody to staphylococcal exotoxins has remained stable in studies from 1960 to 1999 and is correlated with age (see Fig. 64.1 ). The increased incidence of TSS in adolescent girls is related to the lower prevalence of antibody to TSST-1 in this age group. Antibody to TSST-1 in a study of Wisconsin residents was found to be 47% at 1 year of age, 70% at 10 years, 88% at 20 years, and 96% for ages 30 to 50 years during the years 1960 to 1983. A study of women across North America in 1998 to 1999 found similar results, with 85% overall and 81% of subjects aged 13 to 18 years having positive antibody to TSST-1. Transplacental antibody is present in more than 90% of infants.
For the staphylococcal enterotoxins, by 10 years of age, the proportion of individuals with antibody titers of 1 : 100 or higher is 15% for staphylococcal enterotoxin A, 65% for staphylococcal enterotoxin B, 30% for staphylococcal enterotoxin C, 5% for staphylococcal enterotoxin D, and 20% for staphylococcal enterotoxin E. By 22 years of age, the proportion increases to 55% for staphylococcal enterotoxin A, 77% for staphylococcal enterotoxin B, and 98% for staphylococcal enterotoxin C.
Racial, sexual, and geographic differences in antibody prevalence have been noted. Sera randomly selected from 87 control women were seronegative more frequently for antibody to TSST-1 (24%) than were those from 66 control men (9%). A 2003 multicenter study of 3012 women from Ohio, New Jersey, Florida, Arizona, and Manitoba, Canada, found similar rates of S. aureus colonization, but subjects from Manitoba and Arizona were significantly more likely to have a positive antibody titer than were subjects from the other locations.
Interruption of Skin or Mucosal Surface
Most cases of TSS occur in patients with an altered skin or mucosal surface. Trauma or surgery in areas of the body frequently colonized with S. aureus (nose, skin, vagina) places individuals at enhanced risk for infection and subsequent TSS. A wide variety of types of surgeries have been associated with TSS (see Box 64.2 ). Primary deep-tissue staphylococcal infections (e.g., osteomyelitis, pyarthrosis, pyomyositis, endocarditis, renal carbuncles, and bacteremia) are rarely associated with TSS. Types of skin disruptions commonly associated with TSS include burns, surgical incisions, insect bites, needle sticks, tattoos, and varicella. Infected abrasions under casts may be focal sites of TSST-1 production.
Burn wounds provide a particularly rich environment for growth of S. aureus and the production of toxins. *
* References .
In burn centers, TSS occurs predominantly in young children with small burns. In one large pediatric burn center, S. aureus normally was not cultured from any site on admission but was acquired within a few days of admission and became the most common wound pathogen. Sixteen percent of wound isolates of S. aureus produced TSST-1. Only 50% of the children had antibodies to TSST-1 on admission. A toxic shock–like syndrome developed in 13% of children. The mortality rate associated with TSS in children with burns may be as high as 57%.TSS cases following disruption of nasopharyngeal mucosa and the respiratory tract include those associated with sinusitis, pharyngitis, parapharyngeal and submandibular abscesses, tracheitis, pneumonia, rubeola, and influenza. TSS occurring after ear, nose, and throat surgery has been associated with the use of nasal splints and packing materials that may disrupt the ciliary blanket.
The vaginal mucosa may be damaged during placement of tampons or barrier contraceptives, postpartum, or after genital surgery. Heavy colonization of the vagina without any other apparent risk factor has been associated with TSS. Vaginal mucosal infections may provide the right conditions for production of TSST-1, including an aerobic environment, high carbon dioxide concentration, neutral pH, high protein and low glucose concentrations, and low to normal magnesium concentration.
Presence of a Foreign Body
Tampon use was established as a clear risk factor for the development of TSS. Tampons create an aerobic environment in the vagina, which normally is anaerobic. Because TSST-1 production requires oxygen, the increasing concentrations of oxygen may enhance the production of toxin. Tampons may also remove vaginal substrates that normally inhibit the growth of S. aureus. Although it has been suggested that tampons may cause vaginal ulceration leading to greater bacterial growth and toxin absorption, the types of ulcerations seen are likely induced by TSST-1 because they have been found during postmortem examination in women who had never used tampons.
While TSS occurs with all brands of tampons, a greater relative risk for menstrual TSS was found with brands with higher absorbency and particular compositions. In one study, production of TSST-1 varied from undetectable levels to 300 µg/mL depending on the tampon studied. Two studies demonstrated enhanced production of TSST-1 in vitro by the Rely tampon, which was composed of cross-linked carboxymethylcellulose and polyester foam. The polyacrylate rayon and surfactant Pluronic L-92 included in some tampon brands also increase the production of TSST-1 under certain conditions.
In addition to tampons, implanted foreign material, including sutures, central venous lines, metallic or polymeric implants, Teflon splints, and gauze packing, have been documented to enhance the risk for acquiring bacterial infection. These infections are characterized by limited spread beyond the tissues in immediate contact with the implants, poor response to antibiotics, and poor healing without removal of the foreign material. The infection risk varies with the type of foreign material implanted. For example, bacterial adherence is eightfold higher for braided sutures of silk, silicone-heated blue polyester, and absorbable polyglycolic acid than for monofilament nylon. Two important distinctions between TSS and other S. aureus infections related to foreign material are the production of exotoxin and the limited inflammation around the wound.
Other Potential Risk Factors
Other factors that have been examined to identify individuals at increased risk for TSS include personal hygiene practices, human leukocyte antigen type, neutrophil function, adherence of S. aureus to vaginal epithelial cells, alteration of the cervicovaginal flora, and hormonal factors.
Alteration of the vaginal flora, particularly co-colonization with S. aureus and Enterobacteriaceae, has been postulated to enhance the risk for development of TSS. A prospective study found that women who were colonized with toxin-producing S. aureus had higher rates of colonization with Escherichia coli or other Enterobacteriaceae. E. coli isolation rates were 54% in women with TSST-1–positive isolates, 15% in women with TSST-1–negative isolates, and 11% in women with no S. aureus . The significance of these findings is not clear.
Hormonal control is known to be responsible for numerous cyclic changes in vaginal pH and flora. The results of early case-control studies suggested that oral contraceptives might have an effect on vaginal S. aureus organisms that may produce TSS-associated toxins. However, one case-control study found neither protective effect nor enhanced risk associated with oral contraceptive use.
Histopathology
The histopathologic findings on postmortem examination support the concept that TSS is a toxin-mediated disease. Striking histopathologic similarities exist between patients with TSS and those with “scarlet fever” reported in 1936. Typically a total absence of tissue invasion by bacteria and minimal evidence of an inflammatory reaction in most organs are noted. Findings thought to be due to a direct effect of the toxin or mediators (or both) and unrelated to hypoperfusion include subepidermal ulcerations in the cervix, vagina, esophagus, and bladder; depletion of lymphocytes in lymph nodes; a subepidermal cleavage plane in the skin; and mild inflammatory changes in the kidney, liver, heart, and muscle.
Cervicovaginal ulcerations are the only characteristic lesions noted in the genital tract of patients with fatal menstrual TSS, and such ulcers have been found in a patient with menstrual TSS who had never used tampons. The ulcerations are superficial, with separation occurring just beneath the basal layer. Capillary vasodilation and thrombosis with inflammation of the mucosa are present, but no deep-tissue bacterial invasion is seen. The same type of ulcer also has been found in the bladder and esophagus, which suggests that these ulcerations may be caused by the toxins or mediators and not by the use of tampons.
Although the myocardium was described as normal in one postmortem series of TSS, in another series of eight fatal cases, all patients had evidence of focal round-cell infiltration with variable degrees of congestion, edema, and hemorrhage. Myxoid degeneration was found in all heart valves from four patients in one series. Sections of skeletal muscle have demonstrated only congestion, edema, focal hemorrhage or fiber necrosis, and a mild acute inflammatory infiltrate.
Varying degrees of periportal lymphocytic inflammation have been the most consistent findings in the liver; centrilobular congestion with necrosis and mild cellular degeneration also has been described. In the kidney, toxin-mediated mononuclear interstitial nephritis may result from perivasculitis of the adventitia of the renal venules, lesions that probably precede the development of hypotension-induced acute tubular necrosis. The most characteristic findings in the spleen and lymph nodes have been lymphocyte depletion; inactive hypocellular, hypoplastic lymphoid follicles with edema; marked histiocytosis in the interfollicular areas; and hemophagocytosis.
Perivascular lymphocytic infiltrates and bullae that separate at the basement membrane are characteristic of the early skin changes in TSS. No evidence of vasculitis has been reported.
Clinical Spectrum
Acute Phase
The manifestations in TSS are the result of release of endogenous mediators after monocyte and T-cell activation by TSST-1 or the staphylococcal enterotoxins. *
* References .
The physiologic changes are striking in their rapidity of onset and progression and the involvement of almost all body tissues and organs. Endothelial damage with loss of fluid from capillaries into the interstitial space results in loss of peripheral vascular resistance, loss of intravascular volume, and interstitial edema. Prolonged hypotension, interstitial edema, and vascular congestion in turn results in ischemic organ damage. The onset of illness in patients with severe disease is abrupt, with signs and symptoms including fever, chills, malaise, headache, sore throat, myalgia, muscle tenderness, fatigue, vomiting, diarrhea, abdominal pain, and orthostatic dizziness or syncope ( Figs. 64.3 and 64.4 ).During the first 24 to 48 hours, diffuse erythroderma, severe vomiting and watery diarrhea (often with incontinence), decreased urine output, cyanosis, and edema of the extremities may be noted. Some patients may have purpura fulminans. Cerebral ischemia and edema result in somnolence, confusion, irritability, agitation, and occasionally hallucinations, even in individuals without hypotension. Patients with TSS have had signs and symptoms of encephalopathy, cerebral infarction, meningismus, and the cauda equina syndrome.
On initial physical examination, fever, tachycardia, tachypnea, hypotension, erythroderma (generally not seen in patients with severe hypotension or in those without T cells) (see Fig. 64.4A ), and muscle tenderness are noted in conjunction with peripheral cyanosis and edema, conjunctival hyperemia, subconjunctival hemorrhages (see Fig. 64.4B ), beefy red edematous mucous membranes, somnolence, disorientation, and agitation. In menstrual TSS, edema and erythema of the inner aspect of the thighs and the perineum may be noted despite normal findings on uterine and adnexal examination.
In nonmenstrual cases, another focus of infection is present. Surgical wounds and some abscesses colonized or infected with S. aureus and responsible for nonmenstrual TSS often have minimal or no signs of inflammation. The production of TNF-α by macrophages in response to TSST-1 inhibits neutrophil mobilization in vitro, which may provide an explanation for the absence of signs of inflammation. The incubation period for postoperative or postpartum TSS is usually 2 to 4 days but may be as short as 12 hours. Relatively few cases have been associated with deep-tissue infection. Nosocomial acquisition of TSST-1–positive organisms rarely has been documented for postoperative cases. After orthopedic procedures or in association with bone and joint infections, the clinical manifestation of TSS may be confusing as a result of the intense and generalized myalgias associated with TSS, which may be misinterpreted as postoperative musculoskeletal symptoms. The wounds usually appear to be benign.
Laboratory Findings
Laboratory tests reflect the endogenous cytokine release, shock, and organ failure associated with TSS. Leukocytosis may not be present, but the proportion of neutrophils generally exceeds 90%. The proportion of immature neutrophils usually is 25% to 50% of the total number of neutrophils and is associated with absolute lymphopenia. Thrombocytopenia and anemia are present during the first few days and frequently are accompanied by prolonged prothrombin and partial thromboplastin times. Disseminated intravascular coagulation may be present. Sterile pyuria and cerebrospinal fluid pleocytosis are indicative of generalized involvement of the mucous membranes and serosal surfaces. Elevated blood urea nitrogen and creatinine levels reflect kidney damage, and abnormalities in liver function tests reflect liver damage and acute cholestasis. Hypoproteinemia and hypoalbuminemia reflect capillary leak, and profound hypocalcemia may reflect both hypoproteinemia and high serum levels of calcitonin. Muscle involvement is noted by an elevated creatine phosphokinase level, and the hypophosphatemia that occurs despite impaired renal function is unexplained. Most of these test results return to normal within 7 to 10 days of disease onset.
S. aureus is cultured from the cervix or vagina in more than 85% of patients with menstrual TSS and from the focus of infection in patients with nonmenstrual TSS. Positive blood culture results are rare findings. Antibody to TSST-1 or to the staphylococcal enterotoxins is absent at the onset of disease in more than 85% of patients.
Diagnosis
TSS is diagnosed clinically according to its case definition (see Box 64.1 ). Aggressive attempts to find the focus of S. aureus include culture of the cervix and vagina in patients with menses-associated illness and culture of other potentially infected sites in patients with nonmenstrual illness. S. aureus isolates can be examined for their ability to produce TSST-1, although such testing seldom is indicated. This test is of limited usefulness for nonmenstrual cases because TSST-1 is produced by only 40% to 60% of S. aureus isolates from such patients. Isolates from these patients can be examined for the presence of the other enterotoxins. Acute and convalescent sera can be tested for the presence of antibodies to TSST-1 and the other enterotoxins. Elevated levels of anti–TSST-1 in the acute-phase serum of a patient with menstrual-associated TSS is highly unusual.
In most instances, toxin detection tests are of value only for research or for patients with chronic recurrent or atypical disease. Genes for TSST-1 and the enterotoxins have been detected in S. aureus strains by polymerase chain reaction and hybridization techniques. A noncompetitive enzyme-linked immunosorbent assay allows quantitation of TSST-1 in clinical samples. Reversed passive latex agglutination has been used to detect TSST-1 and enterotoxins. Research laboratories also have developed techniques for detecting the selective expansion of V β 2-positive T cells as evidence of host response to superantigenic toxins.
Treatment
The four general principles of treatment of TSS ( Box 64.3 ) are (1) identification and drainage of the focus of toxin production, (2) identification and susceptibility testing of the organism, (3) administration of antimicrobial therapy to block synthesis of the toxin and eradicate the organism, and (4) management of the systemic multiorgan actions of the toxins or mediators. *
* References .
- 1.
Identify the focus of infection: debride and irrigate extensively and remove any foreign material
- 2.
Isolate the organism for antimicrobial susceptibility studies
- 3.
Administer parenteral antimicrobial therapy to stop toxin production and eradicate the organism
- 4.
Manage systemic multiorgan actions of toxins or mediators
- 5.
Administer fluid therapy to maintain adequate venous return and cardiac filling pressure and to prevent end-organ damage
- 6.
Consider intravenous immunoglobulin for the following:
- •
Disease refractory to initial fluid replacement and vasopressor support
- •
A focus of infection that cannot be drained
- •
Location and drainage of the infected site.
The focus of infection should be identified rapidly, any foreign bodies should be removed, and the site should be drained or irrigated completely even if it does not appear to be inflamed. Performing this procedure is of utmost importance because perpetuation of even a small undrained focus of infection may result in serious clinical consequences. If TSS occurs in the immediate postoperative period, the wound should be assumed to be the source of infection regardless of its benign appearance.
Identification and susceptibility testing of the organism.
Isolating the organism for susceptibility testing is of critical importance. The incidence of community-acquired methicillin-resistant S. aureus (MRSA) infections has increased dramatically during the past 10 years. These MRSA strains may be acquired in the community by well individuals with no risk factors and are generally not multiresistant. MRSA strains are fully capable of producing TSST-1. TSS may be caused by either community-acquired or multiresistant hospital-acquired strains of MRSA. To date, vancomycin-resistant strains of S. aureus are rare findings.
Administration of antimicrobial agents.
Administration of antistaphylococcal antibiotics is indicated to stop toxin production and eradicate the organism. The infection may be a superficial or deep-tissue infection and may be associated with bacteremia or bacteriuria. Antistaphylococcal antimicrobial agents should be administered intravenously as soon as possible. Once the patient is stable, high doses of an oral antimicrobial agent to which the organism is susceptible can be given to complete a total course of 10 to 14 days.
Subinhibitory concentrations of the protein synthesis inhibitors clindamycin, erythromycin, clarithromycin, kanamycin, gentamicin, tetracycline, and linezolid have been shown to suppress TSST-1 production in vitro. In one study, clindamycin concentrations the minimal inhibitory concentration (MIC) were effective in totally blocking TSST-1 production.
Data suggest that subinhibitory concentrations of β-lactam antibiotics actually may increase TSST-1 production by S. aureus. At a concentration of half the MIC, nafcillin can increase toxin production 10-fold more than under control conditions. The effect is not seen with vancomycin, another cell wall–active drug, thus suggesting specificity beyond merely a cell wall effect. Coadministration of a protein synthesis inhibitor with a β-lactam antibiotic blocks the effect.
The choice of initial empiric antimicrobial therapy has become more complex as a result of an increase in the number of community-acquired MRSA infections and the spread of multiresistant MRSA strains in hospitals. In the past, the most effective initial empiric therapy was a combination of a β-lactamase–resistant penicillin and clindamycin. With recent concern for MRSA strains causing TSS, many experts recommend initiating therapy with vancomycin and clindamycin. As an alternative for MRSA coverage, linezolid has been used successfully for treatment of staphylococcal TSS. Once the results of susceptibility testing of the organism are available, therapy can be adjusted appropriately.
Management of systemic multiorgan actions of the toxins or mediators
Fluid replacement.
The most important aspect of the nonspecific treatment of symptomatic patients is fluid replacement. Intravascular volume and cardiac filling pressure must be restored rapidly to achieve adequate tissue perfusion. Because of the ongoing capillary leakage, this fluid replacement may far exceed the estimated fluid requirements based on calculated maintenance and fluid deficit volumes. Some adults have required vasopressors and as much as 12 L of fluid during the first 24 hours to stabilize the circulating blood pressure. Pleural, pericardial, and peritoneal effusions and interstitial edema inevitably occur as a result of the continued vascular capillary fluid leak. Close monitoring in an intensive care unit will facilitate treatment of intravascular fluid loss, myocardial dysfunction, hemodynamic derangements, pulmonary edema, acute respiratory distress syndrome, acute renal failure, encephalopathy, and disseminated intravascular coagulation.
Intravenous immunoglobulin and toxin inhibition.
High levels of antibody to TSST-1 and the staphylococcal enterotoxins are present in intravenous immunoglobulin (IVIG) preparations. These antibodies may inhibit the binding of toxins to MHC class II antigen-processing cells or interfere with presentation of toxin by these cells to the T-cell receptor. In animal models of TSS, both monoclonal antibodies to TSST-1 and human IVIG given at the time of inoculation of TSST-1–positive S. aureus prevent the development of TSS. Even when IVIG was administered 29 hours after administration of TSST-1, the increase in survival rates in the IVIG-treated animals still was significant. No adverse reactions were noted in the treated animals, and no evidence was found of disease mediated by the formation of antigen-antibody complexes.
Anecdotal case reports in humans have indicated a beneficial effect in treating staphylococcal TSS. Because IVIG is expensive and most patients respond rapidly once standard therapeutic measures are initiated, some experts reserve IVIG for patients with an inaccessible focus of infection or for those who continue to deteriorate after receiving fluid and vasopressor support (see Box 64.3 ). The dose most often used has been 400 mg/kg given as a single dose over the course of several hours. This dose results in a serum antibody titer of greater than 1 : 100, much higher than that appearing to provide immunity to TSST-1. Some studies have used IVIG doses of up to 2 g/kg. Because early administration of IVIG possibly could blunt the immune response to TSST-1 or other toxins and increase the possibility of a recurrent episode, the potential risks and benefits of this form of therapy must be considered for each patient.
The role of endotoxin in the pathogenesis of TSS is unclear. The failure of polymyxin B and anti-J5 antiserum to alter the course of TSST-1–positive TSS in a rabbit model suggests that endotoxin may not be an important mediator of TSS. Early and sporadic case reports of TSS suggested therapeutic benefit with naloxone, calcium, and exchange transfusion in severely ill patients unresponsive to the usual forms of therapy.
Corticosteroids.
Short courses of methylprednisolone or dexamethasone, if given early in the course of the disease, have been associated with a reduction in the duration of fever and the severity of illness but no reduction in mortality rates. In vitro, dexamethasone has been shown to downregulate TSST-1–induced cytokine production. Because no controlled prospective study has demonstrated efficacy, the use of steroids probably should be restricted to hypotensive patients unresponsive to fluid resuscitation, antimicrobial agents, and IVIG.
Subacute Phase
Once appropriate treatment is initiated, response is usually rapid with defervescence within 48 hours. The hemodynamic changes initially observed include tachycardia, decreased systemic vascular resistance, decreased central venous pressure, hypovolemia, normal pulmonary artery wedge pressure, and an increased cardiac index. Once aggressive fluid therapy has been initiated, myocardial edema and potential failure, along with pulmonary and cerebral edema in the face of renal failure, become the most critical management issues.
The reasons for myocardial failure are unclear but probably are related to perivascular inflammation of the coronary vessels, edema, and postulated myocardial depressant factors. TSST-1 has been shown to inhibit systolic function in isolated rabbit atria, although at higher than usual circulating concentrations. Arrhythmias may result from myocardial damage or electrolyte abnormalities.
During the decompensated stage of myocardial dysfunction, the cardiac index falls and pulmonary wedge pressure increases, with both left atrial and ventricular and diastolic diameters being at the upper limits of normal. Reversible electrocardiographic findings include sinus tachycardia, diffuse loss of voltage, flattened T waves, and diffuse nonspecific ST-T–wave changes. If fatal arrhythmia does not occur during the decompensated stage, the toxic cardiomyopathy is reversible and rarely results in permanent changes. This process is similar to “stunned myocardium,” a transient, postischemic myocardial dysfunctional state.
Pulmonary edema and acute respiratory distress syndrome occur commonly in patients when massive fluid replacement is necessary and capillary leakage continues in the lungs. Pulmonary edema appears rapidly once fluid replacement is initiated and often necessitates intubation and ventilator management for several days.
Forms of TSS-associated acute renal failure include prerenal azotemia and both nonoliguric and oliguric renal failure. The type of renal failure manifested may depend on the degree of intravascular volume depletion. Unless severe acute tubular necrosis necessitates temporary hemodialysis, repletion of intravascular volume usually results in rapid restoration of renal function and, ultimately, diuresis. Permanent renal damage is an extremely rare event.
Management of fluids, electrolytes, and metabolic acidosis in patients with TSS is complex. Although tetany is a rare occurrence, severe hypocalcemia may be life-threatening and should be corrected. Most patients require potassium replacement and management of metabolic acidosis. The use of colloid for fluid replacement and removal of the toxin stimulus for capillary leakage ultimately correct the hypoproteinemia.
The typical dermatologic manifestations follow a predictable sequence (see Fig. 64.3 ). A flaky desquamation begins on the trunk and extremities 5 to 7 days after the onset of symptoms. From days 10 to 12 and for as long as a month, the characteristic full-thickness desquamation of the fingers, toes, palms, and soles takes place ( Fig. 64.5 ). A variety of atypical dermatologic manifestations, including petechiae and subepidermal bullae, have been described. Many patients exhibit desquamation of the mucous membranes, which is particularly painful when the oral mucous membranes are involved. In addition, a small number of patients will have reactivated herpes simplex virus type 1 or 2 lesions with the acute illness. A late-onset pruritic maculopapular rash with edema and low-grade fever unrelated to antimicrobial therapy occurs in more than 50% of menses-associated cases within 7 to 14 days of disease onset. The cause of this late-onset rash is unknown.
The hematologic system seldom is involved with major complications in TSS. Although disseminated intravascular coagulation may be present, gastrointestinal, uterine, or cerebral bleeding rarely occurs. While thrombocytopenia may be present initially in patients with disseminated intravascular coagulation, thrombocytosis is typical in the recovery phase. Mild to moderate normocytic, normochromic anemia, likely due to a combination of hemodilution and suppressed red blood cell synthesis, develops in almost all patients with TSS and resolves during convalescence. Hypoferrinemia occurs commonly.
The relatively common toxic or ischemic encephalopathy, rarely complicated by seizures, resolves slowly during the first 4 to 5 days of hospitalization. The gastrointestinal, musculoskeletal, and hepatic changes resolve rapidly. Sequelae associated with these changes are rare, except for prolonged muscle weakness. Joint manifestations generally are self-limited.
Outcome and Sequelae
Death associated with TSS usually takes place within the first few days of illness, but it may occur as late as 15 days after presentation. Fatalities have been attributed to refractory cardiac arrhythmias, cardiomyopathy, irreversible respiratory failure, and, rarely, bleeding caused by coagulation defects. The duration of circulation of toxins and mediators and the associated hypotension may be the best predictors of the severity of the end-organ damage.
Sequelae that appear to be related to a prolonged period of hypotension include chronic renal failure, gangrene, and telogen effluvium. Hair and nail loss occurs 4 to 16 weeks after onset of the illness, with restoration taking place in 5 to 6 months.
Other sequelae, such as neuropsychological abnormalities, prolonged myalgia and weakness, carpal tunnel syndrome, chronic dermatitis, Raynaud syndrome, and new allergies are explained less easily. Impaired memory and concentration have been found in patients who did not require any therapy other than intravenous fluids to restore their blood pressure. Fertility patterns and pregnancy outcomes were similar in controls and women with TSS, both before and after the illness.
Recurrences
Recurrences for both menstrual and nonmenstrual TSS are well described, usually associated with inadequately treated disease. A small number of patients with menstrual disease and repeated tampon exposure have experienced as many as 6 to 12 recurrences. In most patients the first episode is the most severe. Antistaphylococcal antimicrobial therapy for 10 to 14 days and discontinuation of tampon use reduces the rate of recurrence significantly.
The absent or delayed antibody response to TSST-1 found in both menstrual and nonmenstrual TSS patients probably accounts for the continued susceptibility to TSS and high recurrence rate. The fact that superantigenic toxins are not processed by antigen-processing cells and T lymphocytes as conventional antigens may provide an explanation for the poor convalescent antibody response to these antigens. In mice and one patient with TSS, in vivo TSST-1–induced proliferation was followed by hyporesponsiveness of TSST-1–responsive V β 2 + T cells.
Culture-negative, menses-associated recurrent episodes of TSS continue to be found in a small number of patients despite discontinuation of tampon use and administration of appropriate antimicrobial therapy during the acute episode. Administration of an oral β-lactamase–resistant antistaphylococcal antimicrobial agent during menses has been tried in an attempt to prevent these recurrences but is not always successful. In recurrent cases resistant to this form of prophylaxis, consideration could be given to the empiric use of rifampin, clindamycin, erythromycin, or an oral contraceptive.
Atypical Manifestations
Mild Disease
Recognition of mild episodes of menstrual TSS is particularly important because of the risk of recurrence without appropriate therapy. Patients with mild TSS typically do not form antibody to TSST-1 during convalescence. A mild episode may be recognized only in retrospect, after desquamation, a recurrent episode, or both develop.
The presence of any combination of fever, headache, sore throat, diarrhea, vomiting, orthostatic dizziness, syncope, and myalgia in a menstruating woman or an individual with a potential S. aureus infection should raise suspicion of TSS. A positive culture for S. aureus is helpful but not diagnostic because S. aureus may be cultured from the cervix or vagina in up to 33% of menstruating women. Other laboratory data usually do not reflect multisystem involvement in mild disease, and assays for TSST-1 in body fluids are investigational. Support for the diagnosis of mild TSS depends on the constellation of findings present, including subsequent desquamation of the palms, soles, toes, or fingers; demonstration that S. aureus isolates from the site of infection produce TSST-1 or an enterotoxin; absence of antibody to TSST-1 or enterotoxins in acute-phase serum; and recurrent disease.
Recalcitrant Erythematous Desquamating Disorder
An atypical, subacute variant of TSS has been described in patients with acquired immunodeficiency syndrome (AIDS) and labeled recalcitrant erythematous desquamating disorder. S. aureus strains producing TSST-1 or staphylococcal enterotoxins A or B have been isolated from patients with AIDS presenting with diffuse erythema, extensive cutaneous desquamation, hypotension, and variable multiorgan involvement. The illness is recalcitrant, prolonged, and characterized by multiple recurrences. In one patient, elevated levels of TNF and IL-6 were found during severe episodes. When antitoxin antibodies have been measured during recurrence, they have been undetectable. Two patients responded well to IVIG. TSST-1 and the enterotoxins may exacerbate human immunodeficiency virus type 1 (HIV-1) infection because these toxins have been shown to activate HIV-1 gene expression in vitro. Patients with the combined cellular and humoral immunodeficiencies of AIDS may be at particular risk for the development of severe, frequent, and prolonged episodes of TSS.
Neonatal Toxic Shock Syndrome–like Exanthematous Disease
A toxic shock–like illness related to TSST-1–producing MRSA, labeled neonatal toxic shock syndrome–like exanthematous disease (NTED), was initially described in neonates in Japan and subsequently recognized in France. These infants present in the first week of life with a combination of a generalized erythematous macular rash, thrombocytopenia, elevated acute-phase reactants, and fever. Patients are colonized with a single clonal type of MRSA that produces TSST-1 and staphylococcal enterotoxin C. Although no focus of infection is present and no exotoxins are detectable in serum, analysis of T cells shows the selective expansion of V β 2-positive T cells typically found in response to TSST-1. Maternal IgG antibody to TSST-1 plays a protective role in preventing NTED. Although complications occur in premature neonates, term infants typically recover within 5 days without active treatment. The limited disease has been attributed to the minute amount of exotoxin from colonized sites and the high susceptibility to induction of anergy in the immature T cells of neonates.
Streptococcal Toxic Shock Syndrome
An increase in the incidence and severity of S. pyogenes was recognized in the 1980s. These infections included septicemia, necrotizing fasciitis, and a toxic shock–like syndrome. In 1993, the Working Group on Severe Streptococcal Infections proposed a consensus definition for streptococcal TSS based on clinical criteria including isolation of group A streptococci from a normally sterile site, hypotension, and two or more of the following: renal impairment, coagulopathy, liver dysfunction, acute respiratory distress syndrome, erythematous macular rash, and soft tissue necrosis.
Since the 1990s, the rate of severe streptococcal infections has approximated 2 to 4 cases/100,000 per year. In adults, streptococcal TSS comprises 10% to 15% of severe streptococcal infections, with a mortality rate of 30% to 70%. Streptococcal TSS is less common in children and carries a lower mortality rate than in adults. A multicenter study covering a US population of 29.7 million from 2000 to 2004 identified 5400 cases of invasive group A streptococcal infection (3.5 per 100,000 per year), including 572 children younger than 10 years of age. Of the 572 children with severe streptococcal infection, 26 (4.6%) developed streptococcal TSS with a 7.2% case-fatality rate.
Streptococcal TSS is similar to staphylococcal TSS in that it is mediated by superantigenic toxins and results in endothelial damage, hypotension, and multisystem organ involvement. Both cell-associated and soluble streptococcal virulence factors, including M protein and pyrogenic exotoxins A and B, have been implicated as superantigens mediating the systemic effects of streptococcal TSS. It differs from staphylococcal TSS in numerous respects, including severe generalized hyperesthesia; extreme pain at the site of skin involvement; a slower onset over the course of several days; typical absence of vomiting, diarrhea, and conjunctival injection; and the frequent absence of erythroderma or the presence of only a sandpaper-like rash. In a multinational European study conducted from 2003–04, streptococcal TSS was most often associated with skin and soft tissue infections, particularly cellulitis and necrotizing fasciitis, followed by septic arthritis, bacteremia, sepsis, and meningitis. Notably, 50% of the cases of necrotizing fasciitis were complicated by streptococcal TSS.
Treatment follows the same principles as for staphylococcal TSS: identification of the focus of toxin production, identification and susceptibility testing of the organism, administration of antibiotic therapy to kill the organism and block synthesis of the toxin , and management of the systemic multiorgan actions of the toxins. Although penicillin is the drug of choice for treatment of most group A streptococcal infections, initial broad gram-positive coverage with vancomycin and clindamycin has been advocated given the clinical overlap between streptococcal and staphylococcal TSS. The addition of clindamycin to inhibit protein synthesis is beneficial in cases of streptococcal as well as staphylococcal TSS. In a mouse model of myositis caused by S. pyogenes , clindamycin was more efficacious than penicillin.
IVIG as adjunctive therapy for streptococcal TSS has been supported by in vitro studies and case reports in humans, but clinical studies have been inconclusive. Downregulation of the lymphokine production induced by streptococcal pyrogenic exotoxin A by IVIG has been demonstrated in vitro. A multicenter, randomized, double-blind, placebo-controlled trial to study IVIG for streptococcal TSS in adults showed 3.6-fold higher mortality in the placebo group. However, statistical significance was not achieved because the study identified only 10 IVIG recipients and 11 placebo patients before premature termination due to slow patient recruitment. As with staphylococcal TSS, IVIG has been advocated for cases of streptococcal TSS that are refractory to aggressive therapy or for patients with an inaccessible focus of infection.
Several patients with TSS have been reported to be simultaneously infected with S. aureus and S. pyogenes. Determining which infection primarily was responsible for the manifestations or whether an amplified effect of exotoxins from both organisms was present was not possible.