Abstract
Wilson et al. (Am. J. Surg. 185:369–375, 2003) developed a disease severity classification system for use in complicated skin and skin structure infections (cSSSI). Two phase 3 trials of ceftaroline fosamil in cSSSI provided the opportunity to evaluate the association between Wilson Severity Risk Class and clinical cure rates. Our analyses did not confirm that an association exists between Wilson Severity Risk Class and clinical cure rate and, thus, did not validate its predictive utility.
TEXT
Severity-of-illness scoring systems are valuable for patient management, as evidenced by the widespread adoption in clinical practice of the Pneumonia Patient Outcomes Research Team (PORT) risk classification system for community-acquired pneumonia (1). The PORT system has also been used extensively for risk stratification at enrollment and subsequent analysis of data in clinical trials of new antimicrobials (2, 3).
No comparable, validated, and widely adopted severity-of-illness scoring system exists for complicated skin and skin structure infections (cSSSI), despite the effort by Wilson et al. to develop a scoring system based on clinical trials of cSSSI in the linezolid development program (4). Wilson et al. derived the scoring system from an initial clinical trial data set and then performed a validation in a different cohort of patients. To our knowledge, however, no further validation of their severity scoring system has been performed.
The following prospectively defined exploratory analysis of the Ceftaroline Versus Vancomycin in Skin and Skin Structure Infections (CANVAS) 1 and 2 trials (5, 6) evaluated the relationship between the Wilson cSSSI severity scoring system and clinical cure rates in patients with cSSSI. CANVAS 1 (NCT00424190) and CANVAS 2 (NCT00423657) were phase 3, international, multicenter, randomized, double-blind studies, using identical protocols prospectively designed for pooled analysis, of intravenous ceftaroline fosamil 600 mg every 12 h versus intravenous vancomycin plus intravenous aztreonam for 5 to 14 days in adult patients with cSSSI (7). Randomized patients had a cSSSI of severity substantial enough to require initial hospitalization or treatment in an emergency room and at least 5 days of intravenous antimicrobial therapy, 3 or more local or systemic signs of infection, and infection involving deeper soft tissue or requiring significant surgical intervention or having diabetes mellitus or peripheral vascular disease and cellulitis or abscess in a lower extremity. At the test-of-cure (TOC) visit 8 to 15 days after the last dose of study drug, clinical cure was defined as total resolution of all signs and symptoms of the baseline infection or improvement such that no further antimicrobial therapy was necessary. The primary outcome measure was the clinical cure rate at the TOC visit in both the clinically evaluable (CE) and modified intent to treat (MITT) analysis populations. The CANVAS studies were designed prospectively to capture data that would allow calculation of the Wilson Disease Severity Score, including patient age, four physical examination findings, six laboratory tests, size and degree of infection-site involvement, and presence of a surgical-site infection, bacteremia, or comorbid condition.
Our analysis population included patients in the MITT (randomized patients who received any study drug) population who had complete data for all components of the Wilson Disease Severity Score and had clinical outcome data available. Of 1,396 CANVAS patients, 1,162 (83.2%) had data to calculate the Wilson Disease Severity Score, and 1,079 (77.3%) also had clinical outcome data available. Baseline demographic characteristics, physical examinations, and laboratory parameters were generally comparable for CANVAS patients and the Wilson patients (Tables 1 and 2). CANVAS patients had a higher rate of cardiac/vascular disease, major abscess, deep infection, recovery of Staphylococcus aureus, and large skin lesions (>150 cm2) (Tables 1 and 2). The Wilson Disease Severity Score was calculated for each CANVAS patient and for CANVAS patients pooled across the 2 trials and treatment groups to determine the percentage in each Wilson Severity Risk Class, I through IV. The median Wilson Disease Severity Score for CANVAS patients was 76.5. The possible association between the Wilson Severity Risk Class and the clinical cure rate was tested for significance (chi-squared test, 0.05 significance level). For CANVAS patients, the cure rate did not differ significantly across the four Wilson Severity Risk Class categories (P = 0.241) (Table 3).
Table 1.
Demographic and baseline characteristics
| Characteristic | No. of patients (%), unless otherwise stated |
||
|---|---|---|---|
| Wilson study A: linezolid + oxacillin-dicloxacillin treatment groups (n = 819) | Wilson study B: linezolid + flucloxacillin treatment groups (n = 183) | CANVAS studies: ceftaroline + vancomycin ± aztreonam treatment groups (n = 1,162)a | |
| Sex | |||
| Male | 507 (61.9) | 93 (50.8) | 715 (61.5) |
| Female | 312 (38.1) | 90 (49.2) | 447 (38.5) |
| Mean age ± SD, yrs | 48.0b | 52.1c | 47.9 ± 16.8 |
| Race | |||
| White | 457 (55.8) | 163 (89.1) | 839 (72.2) |
| Black | 118 (14.4) | 18 (9.8) | 84 (7.2) |
| Asian or Pacific Islander | 80 (9.8) | 0 (0.0) | 16 (1.4) |
| Other or unknown | 164 (20.0) | 2 (1.1) | 230 (19.8) |
| Diagnosis | |||
| Cellulitis | 364 (44.4) | 51 (27.9) | 373 (32.1) |
| Skin abscesses | 122 (14.9) | 18 (9.8) | 418 (36.0) |
| Skin ulcer | 29 (3.5) | 10 (5.5) | 92 (7.9) |
| Erysipelas | 81 (9.9) | 30 (16.4) | NAe |
| Infected surgical incision | 51 (6.2) | 30 (16.4) | NA |
| Infected wound | 64 (7.8) | 13 (7.1) | 151 (13.0) |
| Infected bite | 10 (1.2) | 4 (2.2) | 23 (2.0)d |
| Infected burn | NA | NA | 38 (3.3) |
| SSSI on lower extremity in subjects with diabetes mellitus or peripheral vascular disease | NA | NA | 53 (4.6) |
| Other | 92 (11.2) | 27 (14.8) | 14 (1.2) |
| Missing | 1 (0.2) | 0 (0.0) | 0 (0.0) |
| Degree of involvement (deep) | 634 (77.4) | 177 (96.7) | 1,142 (98.3) |
| Clinically relevant pathogens | |||
| Staphylococcus aureus | 283 (34.6) | 51 (27.9) | 731 (62.9) |
| Streptococcus pyogenes | 87 (10.6) | 17 (9.3) | 95 (8.2) |
| Streptococcus agalactiae | 22 (2.7) | 4 (2.2) | 40 (3.4) |
MITT population and subjects with all components of the disease severity score measured (i.e., no missing values).
Mean age ± SD, years: linezolid, 46.7 ± 17.1; oxacillin-dicloxacillin, 49.2 ± 18.5.
Mean age ± SD, years: linezolid, 52.6 ± 19.1; flucloxacillin, 51.7 ± 16.1.
CANVAS studies exclude infections due to human bites.
NA, not applicable or not available.
Table 2.
Disease severity scoring system and patient characteristics at baseline
| Patient characteristic at baseline | Score (points) | No. of patients (%) |
||
|---|---|---|---|---|
| Wilson study A: linezolid + oxacillin-dicloxacillin treatment groups (n = 819) | Wilson study B: linezolid + flucloxacillin treatment groups (n = 183) | CANVAS studies: ceftaroline + vancomycin ± aztreonam treatment groups (n = 1,162)a | ||
| Comorbid condition | ||||
| Cancer | +30 | 26 (3.2) | 10 (5.5) | 24 (2.1) |
| Cardiac | +10 | 37 (4.5) | 6 (3.3) | 215 (18.5) |
| Diabetes | +10 | 141 (17.2) | 14 (7.7) | 205 (17.6) |
| Hepatic | +20 | 21 (2.6) | 4 (2.2) | 107 (9.2) |
| Immunologic | +30 | 11 (1.3) | 1 (0.6) | 42 (3.6) |
| Renal | +10 | 11 (1.3) | 5 (2.7) | 39 (3.4) |
| Respiratory | +30 | 32 (3.9) | 11 (6.0) | 115 (9.9) |
| Transplantationb | +30 | 2 (0.2) | 0 (0.0) | 0 (0.0) |
| Vascular | +10 | 6 (0.7) | 1 (0.6) | 146 (12.6) |
| Physical examination | ||||
| Respiratory rate of >30 breaths/min | +20 | 1 (0.1) | 4 (2.2) | 0 (0.0) |
| Systolic blood pressure of <90 mm Hg | +20 | 0 (0.0) | 0 (0.0) | 3 (0.3) |
| Temperature of <35°C or >40°C | +15 | 4 (0.5) | 2 (1.1) | 2 (0.2) |
| Pulse of >125 beats/min | +15 | 2 (0.2) | 1 (0.6) | 3 (0.3) |
| Laboratory | ||||
| Blood urea nitrogen of >30 mg/dl | +20 | 40 (4.9) | 7 (3.8) | 28 (2.4) |
| Sodium of <130 mEq/liter | +20 | 22 (2.7) | 2 (1.1) | 15 (1.3) |
| Glucose of >250 mg/dl | +10 | 48 (5.9) | 8 (4.4) | 61 (5.3) |
| Hematocrit of <30% | +10 | 41 (5.0) | 3 (1.6) | 34 (2.9) |
| White blood cells count of <4.5 × 103 or >10 × 103 cells/mm3 | +10 | 389 (48.5) | 105 (57.4) | 493 (42.4) |
| Bands of >15% | +10 | 59 (7.2) | 4 (2.2) | 1 (0.1) |
| Others | ||||
| Bacteremia | +20 | 28 (3.4) | 5 (2.7) | 48 (4.1) |
| Degree of involvement (deep) | +10 | 634 (77.4) | 84 (45.9) | 1,142 (98.3) |
| Skin lesion >150 cm2 | +10 | 377 (46.0) | 36 (19.7) | 572 (49.2) |
| Surgical wound infection | +10 | 116 (14.2) | 22 (12.0) | 48 (4.1) |
MITT population and subjects with all components of the disease severity score measured (i.e., no missing values).
Transplant patients were excluded from the CANVAS studies.
Table 3.
Clinical outcome at follow-up by comorbid conditions, disease severity, and risk class
| Classification | Wilson study A |
Wilson study B |
CANVAS studies |
|||
|---|---|---|---|---|---|---|
| Clinical cure,a n/N (%) | P value | Clinical cure,a n/N (%) | P value | Clinical cure,a n/N (%) | P value | |
| Comorbid conditions at baseline | ||||||
| None | 404/487 (83) | 0.023 | 108/129 (84) | 0.139 | 545/585 (93.2) | 0.144 |
| At least 1 | 147/195 (75) | 27/37 (73) | 448/494 (90.7) | |||
| Disease severity score | ||||||
| ≤Medianb | 290/348 (83) | 0.085 | 60/65 (92) | 0.004 | 489/527 (92.8) | 0.431 |
| >medianb | 261/334 (78) | 75/101 (74) | 504/552 (91.3) | |||
| Risk class quartile (from study A) | ||||||
| I (<65 points) | 144/171 (84) | 0.052c | 19/20 (95) | 0.032c | 335/361 (92.8) | 0.241 |
| II (65 to 82 points) | 146/177 (83) | 41/45 (91) | 239/255 (93.7) | |||
| III (83 to 100 points) | 139/168 (83) | 35/46 (76) | 179/194 (92.3) | |||
| IV (>100 points) | 122/166 (74) | 40/55 (73) | 240/269 (89.2) | |||
Excludes patients with missing outcomes (i.e., indeterminates) and only includes subjects with all components of the severity score measured (n = 1,079 for CANVAS studies). n is the number of patients having a clinical outcome of “cured,” while N is the total number of patients having the specified classification.
Studies A and B, median = 82; CANVAS studies, median = 76.5.
P value for comparison of differences among all risk categories.
This primary analysis excluded 234 patients; 18 patients were not dosed with study drug, and the remainder had missing baseline data: hematocrit value (195 patients), band count (151 patients), WBC count (140 patients), glucose value (59 patients), BUN value (20 patients), sodium value (20 patients), respiratory rate (2 patients), systolic blood pressure (1 patient), temperature (1 patient), pulse rate (1 patient), and SSSI lesion area (1 patient) (note: not mutually exclusive reasons). Therefore, the primary reason for not being included in the analysis was unavailability of baseline laboratory values.
Given that the TOC cure rate was so high (>90%) for CANVAS patients (Table 3), additional exploratory analyses were performed for patient outcome rates at earlier time points, including end of treatment (investigator-assessed clinical response) and the U.S. Food and Drug Administration's early response endpoint at treatment day 3 (8) (Table 4). For the latter analyses, success was defined as absence of fever and cessation of SSSI lesion spread, with one analysis performed in all CE patients and a second in the CE subset who had a baseline lesion area of >75 cm2 (8). None of these analyses confirmed any discriminatory ability of the Wilson system.
Table 4.
Clinical outcome at selected time points for CANVAS studies
| Risk class quartile (points) | Success rate, n/N (%)a |
|||
|---|---|---|---|---|
| Test-of-cure clinical response (from Table 3) | End-of-treatment investigator-defined response | Day 3 afebrile and cessation of lesion spread | Day 3 afebrile and cessation of lesion spread (subset with baseline lesion area >75cm2) | |
| I (<65) | 335/361 (92.8) | 360/384 (93.8) | 286/372 (76.9) | 161/229 (70.3) |
| II (65–82) | 239/255 (93.7) | 246/258 (95.4) | 190/253 (75.1) | 134/186 (72.0) |
| III (83–100) | 179/194 (92.3) | 185/196 (94.4) | 132/186 (71.0) | 104/145 (71.7) |
| IV (>100) | 240/269 (89.2) | 254/278 (91.4) | 208/267 (77.9) | 173/225 (76.9) |
| P value | 0.241 | 0.2761 | 0.348 | 0.430 |
n is the number of patients having a clinical outcome of “success,” while N is the total number of patients having the specified risk class quartile.
Recent attempts to define a severity-of-illness scoring system for cSSSI began with a study by Eron et al. in 2000 (9). The Eron system was subsequently modified in an attempt to improve its operational characteristics (10). Marwick et al. (11) published a retrospective analysis of the relationship between disease severity and outcome in a cohort of hospitalized patients with skin and soft tissue infection (SSTI). Their definition of disease severity, which was based on a further modification of the Eron classification, proved capable of identifying differential risk of mortality and inadequate therapy. For example, 30-day mortality and inadequate therapy increased with severity class: I, 1% mortality, 14% inadequate therapy; II, 11% mortality, 39% inadequate therapy; III, 17% mortality, 39% inadequate therapy; and IV, 33% mortality, 92% inadequate therapy. These high mortality rates, compared to those observed in the CANVAS studies, likely reflect the less rigid constraints of a practice setting as opposed to those of a registrational clinical study.
In a prospective analysis, Marwick et al. (12) evaluated 2 disease severity scoring systems (CREST criteria and Ki/Rotstein criteria) and their effect on management in 79 patients with community-acquired SSTIs. Significant discrepancies were identified between the 2 scoring systems, and data describing overtreatment of mild infections and undertreatment of severe infections were emphasized. The authors concluded that the use of specific SSSI severity scores over generic assessments requires further evaluation in a larger prospective study.
When the CANVAS studies were designed and implemented, the potential utility of a well-validated severity-of-illness scoring system was apparent. We decided to add an exploratory analysis that hopefully would allow us to validate the relationship between the Wilson Severity Risk Class and outcome. Neither the prospectively planned TOC analysis nor subsequently performed analyses based on earlier time points confirmed an association between the Wilson Severity Risk Class and the clinical cure rate at TOC in CANVAS patients and, thus, did not validate its predictive utility.
Possible reasons for our failure to validate the Wilson Severity Risk Class are methodological (e.g., inability to classify comorbidities consistently with Wilson) or epidemiological (e.g., differing patient populations, including differing types or rates of nondrug interventions, such as incision and drainage). For the TOC analyses, the lower clinical failure rate in CANVAS versus Wilson patients (∼9% versus ∼18%) may not have provided enough discriminatory variation to allow detection of a trend across the risk classes. However, the consistent lack of discriminatory power across a range of successful outcome rates at various time points mitigates this concern. Further studies are needed to define an accurate, easy-to-use, clinically relevant, and well-validated severity-of-illness scoring system for skin and skin structure infections.
ACKNOWLEDGMENTS
We appreciate the contributions of the CANVAS team and those of David Friedland, Joseph Laudano (who was an employee of Forest Research Institute at the time of the analysis), and Lily Llorens. We also appreciate the thoughtful suggestions of the manuscript reviewers.
G. H. Talbot, T. O'Neal (formerly Baculik), A. Das, and D. Thye were employees of Cerexa, Inc., a wholly owned subsidiary of Forest Laboratories, Inc., New York, NY, at the time the study was designed and initiated.
This study was supported by Forest Laboratories, Inc., which was involved in the design of the study; the collection, analysis, and interpretation of data; and the decision to present these results.
Scientific Therapeutics Information, Inc., provided editorial assistance, which was funded by Forest Research Institute, Inc.
Footnotes
Published ahead of print 5 November 2012
REFERENCES
- 1. Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, Coley CM, Marrie TJ, Kapoor WN. 1997. A prediction rule to identify low-risk patients with community-acquired pneumonia. N. Engl. J. Med. 336:243–250 [DOI] [PubMed] [Google Scholar]
- 2. File TM, Jr, Low DE, Eckburg PB, Talbot GH, Friedland HD, Lee J, Llorens L, Critchley IA, Thye DA, on behalf of the FOCUS 1 investigators 2011. FOCUS 1: a randomized, double-blinded, multicentre, phase III trial of the efficacy and safety of ceftaroline fosamil versus ceftriaxone in community-acquired pneumonia. J. Antimicrob. Chemother. 66(Suppl 3):iii19–iii32 [DOI] [PubMed] [Google Scholar]
- 3. Low DE, File TM, Jr, Eckburg PB, Talbot GH, Friedland HD, Lee J, Llorens L, Critchley IA, Thye D, on behalf of the FOCUS 2 investigators 2011. FOCUS 2: a randomized, double-blinded, multicentre, phase III trial of the efficacy and safety of ceftaroline fosamil versus ceftriaxone in community-acquired pneumonia. J. Antimicrob. Chemother. 66(Suppl 3):iii33–iii44 [DOI] [PubMed] [Google Scholar]
- 4. Wilson SE, Solomkin JS, Le V, Cammarata SK, Bruss JB. 2003. A severity score for complicated skin and soft tissue infections derived from phase III studies of linezolid. Am. J. Surg. 185:369–375 [DOI] [PubMed] [Google Scholar]
- 5. Corey GR, Wilcox M, Talbot GH, Thye D, Friedland D, Baculik T on behalf of the CANVAS 1 investigators 2010. CANVAS 1: the first phase III, randomized, double-blind study evaluating ceftaroline fosamil for the treatment of patients with complicated skin and skin structure infections. J. Antimicrob. Chemother. 65(Suppl 4):iv41–iv51 [DOI] [PubMed] [Google Scholar]
- 6. Wilcox M, Corey GR, Talbot GH, Thye D, Friedland D, Baculik T, on behalf of the CANVAS 2 investigators 2010. CANVAS 2: the second phase III, randomized, double-blind study evaluating ceftaroline fosamil for the treatment of patients with complicated skin and skin structure infections. J. Antimicrob. Chemother. 65(Suppl 4):iv53–iv65 [DOI] [PubMed] [Google Scholar]
- 7. Corey GR, Wilcox M, Talbot GH, Friedland HD, Baculik T, Witherell GW, Critchley I, Das AF, Thye D. 2010. Integrated analysis of CANVAS 1 and 2: phase 3, multicenter, randomized, double-blind studies to evaluate the safety and efficacy of ceftaroline versus vancomycin plus aztreonam in complicated skin and skin-structure infection. Clin. Infect. Dis. 51:641–650 [DOI] [PubMed] [Google Scholar]
- 8. U. S. Food and Drug Administration, Center for Drug Evaluation 3 October 2012, accession date Guidance for industry: acute bacterial skin and skin structure infections: developing drugs for treatment. FDA, Washington, DC: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm071185.pdf [Google Scholar]
- 9. Eron LJ. 2000. Infections of skin and soft tissue: outcomes of a classification scheme. Clin. Infect. Dis. 31:287 [Google Scholar]
- 10. Eron LJ, Lipsky BA, Low DE, Nathwani D, Tice AD, Volturo GA. 2003. Managing skin and soft tissue infections: expert panel recommendations on key decision points. J. Antimicrob. Chemother. 52(Suppl 1):i3–i17 [DOI] [PubMed] [Google Scholar]
- 11. Marwick C, Broomhall J, McCowan C, Phillips G, Gonzalez-McQuire S, Akhras K, Merchant S, Nathwani D, Davey P. 2011. Severity assessment of skin and soft tissue infections: cohort study of management and outcomes for hospitalized patients. J. Antimicrob. Chemother. 66:387–397 [DOI] [PubMed] [Google Scholar]
- 12. Marwick C, Rae N, Irvine N, Davey P. 2012. Prospective study of severity assessment and management of acute medical admissions with skin and soft tissue infection. J. Antimicrob. Chemother. 67:1016–1019 [DOI] [PubMed] [Google Scholar]