Correlation of Histopathologic Findings with Clinical Outcome in Necrotizing Fasciitis

Transcription

1 MAJOR ARTICLE Correlation of Histopathologic Findings with Clinical Outcome in Necrotizing Fasciitis Mohanad Bakleh, 1 Lester E. Wold, 2 Jayawant N. Mandrekar, 3 William S. Harmsen, 3 Haytham H. Dimashkieh, 2 and Larry M. Baddour 1 1 Division of Infectious Diseases, Department of Medicine, 2 Department of Laboratory Medicine and Pathology, and 3 Division of Biostatistics, Mayo Clinic College of Medicine, Rochester, Minnesota Background. Necrotizing fasciitis is an uncommon disease with high morbidity and mortality rates. Little is known about the role of histopathologic examination in disease prognosis. Methods. A retrospective study was conducted to determine what correlations, if any, exist between the histopathologic features of resected tissue in patients with necrotizing fasciitis and clinical outcome. Results. Eighty-two cases of necrotizing fasciitis that occurred between January 1990 and December 2002 were identified. Histopathologic findings were available for review in 63 cases. A novel histopathologic classification scheme, based on hematoxylin-eosin and Gram stain results, was developed. The classification scheme included 3 stages: stage I, characterized by an intense neutrophilic response and an absence of bacteria in infected tissue; stage II, characterized by the presence of a moderate-to-severe neutrophilic response and positive Gram stain results or by minimal to absent neutrophilic response with a negative Gram stain result; and stage III, characterized by the presence of few or no polymorphonuclear leukocytes and a Gram stain result positive for bacteria during histopathologic examination. Patients with stage I findings had a significantly lower risk of death than patients with stage III findings (7.1% vs. 47%; odds ratio [OR], 0.1; 95% confidence interval [CI], ; P p.03). Patients with stage II findings had a significantly lower mortality rate than patients with stage III findings (14.2% vs. 47%; OR, 0.2; 95% CI, ; P p.04). Due to the small number of deaths ( n p 11) in patients for whom histopathologic examination of resected tissue was performed, multivariate analysis was not done. Conclusions. Results of this study suggest that histopathologic findings may correlate with clinical outcome in cases of necrotizing fasciitis. Because the histopathologic scheme is based on results of commonly available stains, it could be easily adopted for use in other institutions that could further evaluate its usefulness as a prognostic tool. Necrotizing fasciitis (NF) is an uncommon, potentially life-threatening infection that involves subcutaneous tissue and causes progressive destruction of fascia and adipose tissue [1, 2]. The reported incidence of NF has increased in recent years [1 5]. The microbiologic causes of NF include mixed aerobic and anaerobic bacteria (type I NF) and b-hemolytic streptococci (type II NF). Early surgical intervention is critical in the management of this fulminant disease [6, 7]. Little is known about the correlation between histologic findings and clinical outcome of NF. There are Received 8 July 2004; accepted 27 September 2004; electronically published 6 January Reprints or correspondence: Dr. Mohanad Bakleh, Dodge City Medical Center, 2020 Central Ave, PO Box 1000, Dodge City, KS (mohanadbakleh@yahoo.com). Clinical Infectious Diseases 2005; 40: by the Infectious Diseases Society of America. All rights reserved /2005/ $15.00 intriguing clinical observations [8 11] and animal model data [12], however, that suggest that there may be prognostic value in histopathologic findings in cases of NF. In both human- and baboon-associated disease, the mortality rate was higher in cases in which there was little neutrophilic response and large concentrations of bacteria in infected tissue. The purpose of this study is to determine whether the findings of histopathologic examination of resected tissue in NF cases have prognostic implications in humans. METHODS This is a retrospective cohort study of patients with a diagnosis of NF who were seen at Mayo Clinic (Rochester, MN) between 1 January 1990 and 31 December Cases were identified by a computer-generated search of medical records using The International Classification of Disease, Ninth Revision. All medical records 410 CID 2005:40 (1 February) Bakleh et al.

2 with a diagnosis of NF, Fournier gangrene, Meleney synergistic gangrene, fasciitis not otherwise specified, soft-tissue infection, infection of the tendon and fascia, or inflammation of tendon and fascia were retrieved and reviewed. Cases that met the definition of NF were included in the study. When 11 surgical intervention was performed, only histopathologic specimens obtained during the initial surgery were considered for study and analysis. Available pathologic specimens of resected tissue obtained from the identified cases were collected. Histopathologic slides were prepared from formalin-fixed, paraffinembedded specimens. Previously stored hematoxylin-eosin stained slides were retrieved, Gram stain was performed, and the results were reviewed. Histopathologic review was performed by a pathologist (L.E.W.) who was blinded to clinical outcomes. The study was approved by Institutional Review Board at Mayo Clinic. Definition of NF. NF was defined by either histopathologic or surgical findings [2, 7]. Histopathologic evidence of NF included the presence of necrosis of superficial fascia (with or without polymorphonuclear infiltrates or bacteria) and edena of the reticular dermis, subcutaneous fat, and superficial fascia. In the absence of findings of pathologic examination of resected tissue, the diagnosis required characteristic surgical findings that included the presence of gray necrotic fascia with lack of resistance to blunt dissection, absence of bleeding during surgical dissection, and presence of foul-smelling dishwater pus at surgery in type I NF cases. In addition to the above criteria, the patient was also required to have moderate-to-severe sepsis judged to be secondary to necrotizing fasciitis. Data collection. Data collection included age, sex, predisposing and associated conditions, site of the infection, pertinent laboratory findings, blood and tissue culture results, type of antibiotics used, timing and number of surgeries after hospital admission (or after the onset of symptoms in nosocomial cases), duration of hospitalization, prior use of nonsteroidal antiinflammatory drugs (NSAIDs), clindamycin and intravenous immunoglobulin use in cases of b-hemolytic streptococcal infection, and the diagnosis of streptococcal toxic shock syndrome in cases caused by b-hemolytic streptococci. Severity of illness was measured by calculating the highest APACHE II score within 24 h after hospital admission [13]. Clinical outcome was classified into 1 of 3 categories: (1) inhospital survival, no major sequelae, and limited surgical debridement; (2) in-hospital survival, amputation of 2 fingers, orchiectomy in Fournier gangrene, or failure of 1 organ within 1 week after hospitalization; and (3) in-hospital death. Only short-term (i.e., 28-day) in-hospital death or survival was analyzed. Organ failure was defined by the presence of 1 of the following: respiratory failure requiring mechanical ventilation; hypotension requiring vasopressors; renal failure with a serum creatinine level of 1.6 the upper limit of normal (or 2 the baseline level in patients with chronic renal insufficiency); hepatic injury with a transaminase level of 2 times the upper limit of normal (or 2 the baseline level in patients with chronically elevated levels); coagulation abnormalities with abnormal prothrombin time, partial thromboplastin time, or platelet count!100,000 platelets/ml; and encephalopathy with altered mental status. Streptococcal toxic shock syndrome was defined according to criteria established by the Working Group on Severe Streptococcal Infection [14]. Data on histopathologic inflammatory response and Gram stain results were combined in a staging scheme, illustrated in table 1. Three grades of inflammatory response in infected tissue were defined: (1) few neutrophils (including a single neutrophil or no neutrophils); (2) moderate neutrophilic presence in small groups or clusters of neutrophils without associated necrotic changes; and (3) severe neutrophilic infiltration with microabscess formation and necrosis. Gram stain was performed on all available pathologic tissue blocks, and the results were reported as either positive or negative, depending on whether bacteria were seen or not. Statistical methods. Association between histopathologic findings (stage I, II, or III) and clinical outcome was the primary analysis. The primary event of interest was clinical outcome, including in-hospital death, amputation, and organ failure at 1 week after hospital admission. We conducted a univariate logistic regression analysis to statistically determine whether there was an association between specific demographic and clinical variables and outcomes that included amputation, organ failure, and death. Variables included age, sex, site of infection, number and types of comorbidities, blood and tissue culture results, timing of debridement surgery after hospital admission, number of operative debridements, duration of hospitalization, prior use of nonsteroidal antiinflammatory drugs, type of antibiotics administered, clindamycin and intravenous immunoglobulin use in b-hemolytic streptococcal cases, APACHE II score within 24 h after admission, and histopathologic stage. Multiple variable logistic regression was not done because of the relatively small number of in-hospital deaths. Table 1. Histopathologic stage Stage I Stage II Stage III Histopathologic staging system of necrotizing fasciitis. Characteristics Moderate-to-severe neutrophilic infiltration with no bacteria in the infected tissue Moderate-to-severe neutrophilic presence with bacteria in infected tissue, or few or no neutrophils with no bacteria Few or no neutrophils with bacteria present in infected tissue Histopathologic Features of Necrotizing Fasciitis CID 2005:40 (1 February) 411

3 RESULTS Medical records of 406 potential cases were reviewed, and 82 cases met the case definition. Eighty-two episodes of NF in 81 patients occurring from January 1990 through December 2002 were identified. Forty-six cases (57%) had histopathologic findings typical of NF that met the case definition. The remaining 35 cases (43%) either did not have histopathologic specimens available for review (25 cases) or had histopathologic findings that were nondiagnostic (10 cases), but they all had surgical findings typical of NF that met the case definition. The initial episodes for each of the 81 patients were analyzed. Demographic characteristics and clinical data are summarized in table 2. Results of bacteriologic analysis of identified cases are illustrated in table 3. Blood cultures were positive for bacteria in 15 (20%) of 75 cases, whereas tissue cultures taken from first surgical debridement were positive for bacteria in 70 (96%) of 73 examined cases. Infections due to b-hemolytic streptococci were the most common (31 cases), followed by polymicrobial infections (23 cases) and infections due to Staphylococcus aureus (6 cases). In cases of polymicrobial infection, Enterococcus species, viridans group streptococci, S. aureus, Escherichia coli, and anaerobes were the most common isolates. APACHE II score in the first 24 h of hospitalization was used to measure severity of illness. The mean score was APACHE II score strongly correlated with mortality rate, as expected ( P!.001). An increase of 1 point in the APACHE II score was associated with an increase in the odds of death of 20% (OR, 1.2; 95% CI, ). Surgery was performed as soon as the diagnosis of necrotizing fasciitis was suspected. Ninety-one percent of patients had surgical intervention within either the first 24 h (48 [60%] of the patients) or 48 h (25 [31%] of the patients) of hospitalization. The mean time between hospital admission and initial surgery was 0.6 days (range, 0 6 days). The time between hospital admission and surgery was comparable for the 3 histopathologic stages (56% of stage I cases, 57% of stage II cases, and 67% of stage III cases had surgery in the first 24 h of hospitalization). NF caused considerable morbidity and mortality. Twentyfour (30%) of the patients underwent amputation, and 29 (36%) developed failure of 1 organ after 1 week of hospitalization. Sixteen (20%) of the patients died within 28 days after admission to the hospital. Sixty-three fascia and muscle biopsy specimens were available for review. Hematoxylin-eosin stain of histopathologic specimens was available in 55 cases, and Gram stain was available in 62 cases. Gram stain was positive in 29 (47%) of the cases. An NF staging system that combines both hematoxylin-eosin stain and Gram stain findings was developed (table 1). There was a significant correlation between histopathologic stage and mortality rate in 51 patients with available histopathologic staging data (figure 1). Patients with stage I findings had a significantly lower risk of death than did patients with stage III findings (7.1% vs. 47%; OR, 0.1; 95% CI, ; P p.03). Patients with stage II findings had a significantly lower mortality rate than patients with stage III findings (14.2% vs. 47%; OR, Table 2. Demographic and clinical characteristic of patients with necrotizing fasciitis. Patients with necrotizing fascitis Variable (n p 81) Sex Male 57 (70) Female 24 (30) Age, years Range Mean SD Comorbidities Diabetes mellitus 28 (35) Chronic systemic disease a 17 (22) Age 165 years 22 (27) Surgery 15 (19) Malignancy b 8(9) Compromised immune system c 5(6) Blunt trauma 7 (9) Alcohol abuse 6 (7) Penetrating trauma 4 (5) Chronic skin condition d 3(4) Site of infection Lower extremity- 25 (31) Genital area- 23 (28) Upper extremity- 19 (23) Trunk- 9 (11) Neck- 5 (6) Surgery Time performed after admission, days Range 0 6 Mean SD 0.6 (1.1) Mean no. per patient (range) 4.1 (0 17) Length of hospitalization, days Range Mean SD APACHE II score Range 0 32 Mean SD Clinical outcome: Death 16 (20) Amputation 24 (30) Failure of 1 organ 1 week after admission 29 (36) NOTE. Data are no. (%) of patients, unless otherwise indicated. a Chronic cardiac disease, peripheral vascular disease, pulmonary disease, or renal disease. b Includes only malignancies receiving therapy. c As a result of collagen vascular disease, corticosteroid use, receipt of an organ transplant, HIV infection, or asplenia. d Decubitus/ischemic ulcer or psoriasis. 412 CID 2005:40 (1 February) Bakleh et al.

4 Table 3. Results of blood and resected tissue cultures for patients with necrotizing fasciitis. No. (%) of positive Type of culture, pathogen isolated culture results Blood a (n p 75) b-hemolytic streptococci 8 (11) Staphylococcus aureus 2(3) Streptococcus pneumoniae 1(1) Escherichia coli 1(1) Pseudomonas aeruginosa 1(1) Clostridium perfringens 1(1) Clostridium cadaveris 1(1) Tissue b (n p 73) b-hemolytic streptococci 31 (42) Polymicrobial 23 (31) S. aureus 6 (8.1) Clostridium septicum 3(4) Other anaerobes 2 (3) S. pneumoniae 1(1) Enterococcus species 1 (1) E. coli 1(1) Corynebacterium species 1 (1) Coagulase-negative staphylococci 1 (1) a Sixty (80%) of 75 blood cultures showed no growth. b Three (5.4%) of 73 tissue cultures showed no growth. 0.2; 95% CI, ; P p.04). There was no significant association between histopathologic stage and other clinical outcomes, including amputation ( P p.7) and organ failure (P p.27). When results of Gram stain were excluded, the degree of acute inflammation remained strongly correlated with mortality. Patients with absent or minimal acute inflammation in infected tissue had a significantly higher mortality rate than did patients with moderate to severe acute inflammation (40.7% vs. 3.5%; OR, 0.07; 95% CI, ; P p.02). Among all demographic, clinical, and histopathologic variables that were studied, only APACHE II score and histopathologic classification had a statistically significant correlation with mortality rate ( P!.001). The number of patients with stage III histopathology was too small to statistically compare the shortterm mortality rates for patients with type I and type II NF (4 of 10 patients vs. 3 of 5 patients). Many of the 31 patients who had b-hemolytic streptococcal infection had severe disease. Sixteen patients developed streptococcal toxic shock syndrome. Ten cases required amputation, and 9 patients died. Among patients with b-hemolytic streptococcal NF, those with streptococcal toxic shock syndrome had a 50% mortality rate, and those without had a 6.7% mortality rate (OR, 14; 95% CI, ; P p.02). Patients with b-hemolytic streptococcal infection who were given clindamycin had a significantly lower mortality rate than did those who did not receive clindamycin (14% vs. 60%; OR, 0.13; 95% CI, ; P p.03). DISCUSSION This is, to our knowledge, the first published investigation that examined the prognostic value of histopathologic findings of resected tissue in cases of NF. We found that the degree of acute inflammatory response correlated with clinical outcome. Gram stain results strengthened this correlation with the short-term mortality rate. Histopathologic evaluation can be helpful in both establishing the diagnosis of NF and potentially predicting clinical outcome of the disease. Although a large sample size of NF cases was included in our study, there was an insufficient number of in-hospital deaths (the event of interest) to allow creation of a statistical model. Thus, confounding and interaction among variables identified by univariate analysis could not be examined. We were interested in conducting this investigation based on previous observations at our institution [8] and at others [9, 10], coupled with findings in a baboon model of NF [12]. Cockerill et al. [8] described an outbreak of 16 cases of invasive group A streptococcal disease in patients who presented to Mayo Clinic in A striking and unexpected lack of acute inflammatory cells in the presence of numerous streptococci was found in superficial subcutaneous tissues in 4 patients with deep-seated tissue infection. Three of the 4 patients had NF, and all 3 patients died. Sentinel pathogenesis studies using an animal model of NF have demonstrated a correlation between histopathologic findings and clinical outcome [12]. Thirteen baboons were inoculated intramuscularly with group A streptococci as a primate model of NF and myositis. In the 2 animals that died, the site of bacterial inoculation exhibited intravascular aggregate of neutrophils but no influx of neutrophils in areas with extensive bacterial infection. In the remaining 11 animals, an intense neutrophilic influx without bacteria was seen. Figure 1. Mortality rate based on histopathologic stage of necrotizing fasciitis. P p.03 in a comparison between stage I and stage III (OR, 0.1; 95% CI, ). P p.04, in a comparison between stage II and stage III (OR, 0.2; 95% CI, ). Histopathologic Features of Necrotizing Fasciitis CID 2005:40 (1 February) 413

5 Pathogenic mechanisms that account for the lack of acute inflammatory response in infected NF tissue remain to be elucidated. These mechanisms may involve both host and pathogen factors. Host-related genetic factors that control release of cytokines that promote or inhibit acute inflammation in cases of infection may be operative [15]. Human leukocyte antigen class II allelic variation may contribute to differences in the severity of infection due to b-hemolytic streptococci. Microbial virulence factors that may impact the acute inflammatory response include both gram-positive and gram-negative bacterial products [16 21]. In particular, a-toxin of S. aureus, endotoxins, and type III antigen of group B streptococci can inhibit neutrophil chemotaxis. S. aureus superantigens can indirectly cause accelerated death of neutrophils [16]. The observation that clindamycin use was associated with a reduction in mortality rate among patients with b-hemolytic streptococcal infection in our study has been previously described in other work [22, 23]. Although this observation is consistent with findings from previous investigations, specific treatment details would need to be collected and analyzed to determine the benefit of clindamycin, and our study was not designed to examine treatment efficacy. Limitations of the current investigation include a retrospective study design, which introduces potential bias. NF is an uncommon disease, and it is difficult to conduct a prospective evaluation within a reasonable time. In summary, a histopathologic classification scheme based on routinely available staining procedures was used to predict mortality for patients with NF. Paucity of acute inflammatory response and the presence of bacteria at the site of infection were associated with a poorer clinical outcome. Further studies are needed to confirm our initial findings. Acknowledgments We thank Drs. Frank Cockerill and Kathy Haden for their insightful discussions and Kathy Parsons for assistance with patient chart access. Financial support. This investigation was funded in part by Mayo Clinic College of Medicine. Potential conflicts of interest. All authors: no conflicts. References 1. Bisno AL, Stevens DL. Streptococcal infections of skin and soft tissues. N Engl J Med 1996; 334: Kaul R, McGeer A, Low DE, Green K, Schwartz B. Population-based surveillance for group A streptococcal necrotizing fasciitis: clinical features, prognostic indicators, and microbiologic analysis of seventyseven cases. Ontario Group A Streptococcal Study. Am J Med 1997; 103: 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: Martin PR, Hoiby EA. Streptococcal serogroup A epidemic in Norway Scand J Infect Dis 1990; 22: Sharkawy A, Low DE, Saginur R, et al. Severe group a streptococcal soft-tissue infections in Ontario: Ontario Group. Clin Infect Dis 2002; 34: McHenry CR, Piotrowski JJ, Petrinic D, Malangoni MA. Determinants of mortality for necrotizing soft-tissue infections. Ann Surg 1995; 221: Wong CH, Chang HC, Pasupathy S, Khin LW, Tan JL, Low CO. Necrotizing fasciitis: clinical presentation, microbiology, and determinants of mortality. J Bone Joint Surg Am 2003; 85A: Cockerill FR III, Thompson RL, Musser JM, Schlievert PM, Talbot J, Holley KE. Molecular, serological, and clinical features of 16 consecutive cases of invasive streptococcal disease. Southeastern Minnesota Streptococcal Working Group. Clin Infect Dis 1998; 26: Wlodaver CG. Minimal inflammatory response in some cases of group A streptococcal myofasciitis. J Okla State Med Assoc 1995; 88: Case Records of the Massachusetts General Hospital. Weekly clinicopathological exercises. Case A 33-year-old man with a sore throat followed by swelling and pain. N Engl J Med 1995; 333: Stevens DL. Streptococcal infections of skin and soft tissues. In: Mandell GL, Stevens DL, eds. Atlas of infectious diseases. Philadelphia: Current Medicine, 1995: Taylor FB Jr, Bryant AE, Blick KE, et al. Staging of the baboon response to group A streptococci administered intramuscularly: a descriptive study of the clinical symptoms and clinical chemical response patterns. Clin Infect Dis 1999; 29: Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: Defining the group A streptococcal toxic shock syndrome: rationale and consensus definition. The Working Group on Severe Streptococcal Infections. JAMA 1993; 269: Kotb M, Norrby-Teglund A, McGeer A, et al. An immunogenetic and molecular basis for differences in outcomes of invasive group A streptococcal infections. Nature Med 2002; 8: Schuberth HJ, Krueger C, Hendricks A, Bimczok D, Leibold W. Superantigen-dependent accelerated death of bovine neutrophilic granulocytes in vitro is mediated by blood mononuclear cells. Immunobiology 2000; 202: Schmeling DJ, Gemmell CG, Craddock PR, Quie PG, Peterson PK. Effect of staphylococcal alpha-toxin on neutrophil migration and adhesiveness. Inflammation 1981; 5: Bignold LP, Rogers SD, Siaw TM, Bahnisch J. Inhibition of chemotaxis of neutrophil leukocytes to interleukin-8 by endotoxins of various bacteria. Infect Immun 1991; 59: Bryant RE, Des Prez RM, Rogers DE. Effects of bacterial endotoxin on leukocyte migration, adhesiveness, and aggregation. Studies on human leukocyte motility. Vol. II. Yale J Biol Med 1967; 40: McFall TL, Zimmerman GA, Augustine NH, Hill HR. Effect of group B streptococcal type-specific antigen on polymorphonuclear leukocyte function and polymorphonuclear leukocyte-endothelial cell interaction. Pediatr Res 1987; 21: Stevens DL, Tweten RK, Awad MM, Rood JI, Bryant AE. Clostridial gas gangrene: evidence that alpha and theta toxins differentially modulate the immune response and induce acute tissue necrosis. J Infect Dis 1997; 176: 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: Mascini EM, Jansze M, Schouls LM, Verhoef J, Van Dijk H. Penicillin and clindamycin differentially inhibit the production of pyrogenic exotoxins A and B by group A streptococci. Int J Antimicrob Agents 2001; 18: CID 2005:40 (1 February) Bakleh et al.