SPINE Volume 40, Number 2, pp E112-E119 ©2015, Lippincott Williams & Wilkins
SURGERY
Risk Factors for Surgical Site Infections After Pediatric Spine Operations Lindsay D. Croft, MS,* Jean M. Pottinger, MA, RN,† Hsiu-Yin Chiang, PhD, MS,‡§ Christine S. Ziebold, MD, PhD, MPH,¶ Stuart L. Weinstein, MD, and Loreen A. Herwaldt, MD, FIDSA†‡§
Study Design. Matched case-control study. Objective. To identify modifiable risk factors for surgical site infections (SSIs) after pediatric spinal fusion. Summary of Background Data. The number of SSIs after pediatric spinal fusions increased. Methods. Between July 2001 and July 2010, 22 of 598 pediatric patients who underwent spinal fusion at a university hospital acquired SSIs. Each patient with an SSI was matched with 2 controls by procedure date. Bivariable and multivariable analyses were used to identify risk factors for SSIs and outcomes of SSIs. Results. Gram-negative organisms caused more than 50% of the SSIs. By multivariable analysis, neuromuscular scoliosis (odds ratio [OR] = 20.8; 95% confidence interval [CI], 3.1–889.5; P < 0.0001) and weight-for-age at the 95th percentile or higher (OR = 8.6; 95% CI, 1.2–124.9; P = 0.02) were preoperative factors associated with SSIs. Blood loss (OR = 1.0; 95% CI, 1.0–1.0; P = 0.039) and allografts and allografts in combination with other grafts were operative risk factors for SSIs. The final overall risk model for SSIs was weight-forage at the 95th percentile or higher (OR = 4.0; 95% CI, 1.4–⬁; P = 0.037), American Society of Anesthesiologists score 3 or more (OR = 3.8; 95% CI, 1.6–⬁; P = 0.01), and prolonged operation duration (OR = 1.0/min increase; 95% CI, 1.0–1.0; P = 0.004). SSIs were associated with 2.8 days of additional postoperative length of stay (P = 0.02). Neuromuscular scoliosis was the only factor significantly associated with hospital readmission (OR = 23.6; 95% CI, 3.8–147.3; P = 0.0007). From the *Department of Epidemiology and Public Health, University of Maryland School of Medicine, Baltimore; †Clinical Quality, Safety, and Process Improvement, University of Iowa Hospitals and Clinics, Iowa City; ‡Department of Epidemiology, University of Iowa College of Public Health, Iowa City; and Departments of §Internal Medicine, ¶Pediatrics, and Orthopedics, University of Iowa Carver College of Medicine, Iowa City. Acknowledgment date: March 19, 2014. First revision date: September 25, 2014. Second revision date: October 14, 2014. Acceptance date: October 29, 2014. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. Relevant financial activities outside the submitted work: travel/ accommodations/meeting expenses, board membership, grants. Address correspondence and reprint requests to Loreen A. Herwaldt, MD, FIDSA, Department of Internal Medicine, The University of Iowa Hospitals and Clinics, 200 Hawkins Dr, Iowa City, IA 52242; E-mail: loreen-herwaldt@ uiowa.edu DOI: 10.1097/BRS.0000000000000693
E112
www.spinejournal.com
Conclusion. Our results suggest that pediatric patients undergoing spinal fusion might benefit from antimicrobial prophylaxis that covers gram-negative organisms. Surgical duration, graft implantation, and blood loss are potentially modifiable operative risk factors. Neuromuscular scoliosis, high weight-for-age, and American Society of Anesthesiologists scores 3 or more may help surgical teams identify patients at high risk for SSI. Key words: surgical site infection, spinal fusion, pediatric, complication, scoliosis. Level of Evidence: 4 Spine 2015;40:E112–E119
S
urgical site infections (SSIs) after pediatric spine operations can have devastating consequences. A report from the Scoliosis Research Society indicates that the rate of SSIs for patients younger than 21 years undergoing spine operations is 1.4% for patients with idiopathic scoliosis, 2.2% for patients with congenital scoliosis, 5.5% for patients with neuromuscular scoliosis, and 5.7% for patients with post-traumatic scoliosis.1 Patients with myelomeningocele have the highest reported SSI rates after pediatric spine fusion procedures.2,3 Previous studies have identified possible risk factors both for SSIs1,4–12 and for major complications, including malnutrition,8 cognitive delay,10 absence of drains,7 blood loss,9 blood transfusions,7 cell count,11 use of allografts rather than autografts,10 a combined anterior and posterior surgical approach,12 and the underlying diagnosis.7,9 Most studies before 2000 were case series that lacked comparison groups, were too small for statistical analysis of risk factors,13–15 or included both adults and children and did not examine pediatric cases separately.6,9,16 Moreover, some studies included spinal procedures other than fusions17 and some assessed risk factors for all surgical complications.9,12,17–20 Some studies of pediatric populations evaluated only patients with specific operative indications, such as idiopathic scoliosis7,18,21,22 or neuromuscular scoliosis.8,11 Results of population-specific studies may not be generalizable to a broader population of children undergoing spinal fusions. The baseline SSI rate after pediatric spinal fusions at the University of Iowa Hospitals and Clinics was 2.59% during July 2001 to December 2006 and the rate increased to 3.92% during January 2007 to September 2009. Given increased SSI rates and gaps in our understanding of risk factors for SSIs in January 2015
Copyright © 2015 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE140399_LR E112
23/12/14 10:53 PM
SURGERY
Pediatric SSI Post–Spinal Fusion • Croft et al
these patients, we conducted a case-control study to identify modifiable risk factors.
RESULTS
MATERIALS AND METHODS
During the study period, 598 children (mean age = 13 yr) underwent 615 spinal fusions; 22 children acquired SSIs (3.6%). The mean time to onset of SSIs was 46 days (median = 26 d; range = 3–352 d). Eleven SSIs were superficial incisional and 11 were deep incisional infections. Nine of 11 patients requiring reoperations for SSIs had deep incisional infections. Wound samples from 16 case patients (72%) were cultured and all showed positive results. Of the 16 positive cultures, 10 (62.5%) cultures grew more than 1 organism, 12 (75%) grew at least 1 gram-negative organism, and 11 (69%) grew at least 1 gram-positive organism (Table 1). Five cultures grew Staphylococcus aureus; 1 isolate was resistant to methicillin. Four wound cultures grew coagulase-negative staphylococci, 3 grew Enterobacter cloacae, and 5 grew Pseudomonas aeruginosa. Six superficial SSI samples were not cultured.
We included pediatric spine procedures performed by 1 orthopedic surgeon at the University of Iowa Hospitals and Clinics during July 1, 2001, to July 31, 2010. We defined cases as patients who acquired infections meeting the Centers for Disease Control and Prevention’s National Health Safety Network’s definition of SSI.23 Twenty-two of 598 pediatric patients undergoing spinal fusions during the study period acquired SSIs. We matched each case with 2 patients without SSIs (controls): the patient who underwent spinal fusion immediately before the case and the patient who underwent spinal fusion immediately after the case. We abstracted demographic data (e.g., age, sex, and insurance status) and data on numerous variables from the patients’ medical records: preoperative weight, height, body mass index, underlying diseases, medications, prior spine operations, preoperative laboratory values (hemoglobin, lymphocyte count, glucose [minimum and maximum within 24 hr before the operation and 48 hr after the operation], hemoglobin A1c, albumin, urine analysis, and urine culture), length of hospital stay, the American Society of Anesthesiologists (ASA) score, surgical indication, surgical approach, instrumentation type, bone graft type, blood loss, transfusions, drain placement, lowest intraoperative temperature, lowest intraoperative fraction of inspired oxygen, procedure duration, wound culture results, antimicrobial therapy, seroma formation, hematoma formation, and reoperations. Data were entered into a Microsoft Excel spreadsheet and analyzed using SAS 9.2 (SAS Institute, Cary, NC). We used the Fisher exact test to analyze categorical variables and the Student t test to analyze continuous variables. If data on comorbidities, preoperative medication, antimicrobial prophylaxis, or transfusion were not available, we assumed that the patient did not have the comorbidities or did not receive the treatment. We constructed preoperative, operative, and overall (included patient-related and procedure-related factors) multivariable models to identify factors associated with SSIs. We included variables without missing values and with P ≤ 0.10 or that we presumed to have biological significance in multivariable ordinary logistic regression models. We used stepwise selection to construct an overall multivariable model; P ≤ 0.15 was required for both entering and remaining in the final model. After we identified the final models, we used conditional logistic regression to adjust for the clustering effects of matched case patients and controls. We used linear regression to identify variables that influenced postoperative length of stay and an ordinary logistic regression model to identify variables predicting readmission. We assessed the predictive power of each ordinary logistic regression model with the c-statistic, where 0.5 and 1.0 correspond to no predictive power and perfect predictive power, respectively. We used the Hosmer-Lemeshow goodness-of-fit test to assess how well the models fit the data. Variables with P values of less than 0.05 were considered statistically significant. Spine
Surgical Site Infections
Bivariable Analyses Case patients and controls did not differ significantly by age, height, weight, most comorbidities, and numerous surgical
TABLE 1. Combinations of Organisms Cultured
From Surgical Site Infections
Organism
No. Case Patients (N = 22)
Bacteroides fragilis, Enterobacter cloacae, Staphylococcus aureus
1
B. fragilis, mixed anaerobes including Proteus vulgaris
1
CNS
2
CNS, Haemophilus parainfluenzae, α-hemolytic streptococci
1
E. cloacae
1
Escherichia coli, Pseudomonas aeruginosa
1
Enterococcus spp., Proteus mirabilis
1
Enterococcus spp., E. coli, Klebsiella pneumoniae, CNS
1
Enterococcus spp., K. pneumoniae, P. aeruginosa (2 strains)
1
MRSA, Propionibacterium acnes
1
P. aeruginosa
1
P. aeruginosa (2 strains)
1
S. aureus
1
S. aureus, E. cloacae
1
S. aureus, P. aeruginosa (2 strains)
1
No culture
6
CNS indicates coagulase-negative staphylococci; MRSA, methicillin-resistant Staphylococcus aureus.
www.spinejournal.com
E113
Copyright © 2015 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE140399_LR E113
23/12/14 10:53 PM
SURGERY
Pediatric SSI Post–Spinal Fusion • Croft et al
TABLE 2. Bivariable Analysis* Patients With SSIs (N = 22)
Control Patients (N = 44)
Odds Ratio (95% CI)
P
13.1 ± 3.7
13.4 ± 3.0
…
0.73
Male sex
11 (50.0)
16 (36.4)
1.8 (0.6–4.9)
0.30
No private insurance
3 (13.6)
11 (25.0)
0.5 (0.1–1.9)
0.35
144.4 ± 25.3
153.9 ± 19.1
…
0.09
53.2 ± 26.4
49.0 ± 21.8
…
0.49
8 (36.4)
2 (4.6)
12.0 (2.3–63.3)
0.002
23.9 ± 5.9
19.1 ± 2.9
…
SURGERY
Risk Factors for Surgical Site Infections After Pediatric Spine Operations Lindsay D. Croft, MS,* Jean M. Pottinger, MA, RN,† Hsiu-Yin Chiang, PhD, MS,‡§ Christine S. Ziebold, MD, PhD, MPH,¶ Stuart L. Weinstein, MD, and Loreen A. Herwaldt, MD, FIDSA†‡§
Study Design. Matched case-control study. Objective. To identify modifiable risk factors for surgical site infections (SSIs) after pediatric spinal fusion. Summary of Background Data. The number of SSIs after pediatric spinal fusions increased. Methods. Between July 2001 and July 2010, 22 of 598 pediatric patients who underwent spinal fusion at a university hospital acquired SSIs. Each patient with an SSI was matched with 2 controls by procedure date. Bivariable and multivariable analyses were used to identify risk factors for SSIs and outcomes of SSIs. Results. Gram-negative organisms caused more than 50% of the SSIs. By multivariable analysis, neuromuscular scoliosis (odds ratio [OR] = 20.8; 95% confidence interval [CI], 3.1–889.5; P < 0.0001) and weight-for-age at the 95th percentile or higher (OR = 8.6; 95% CI, 1.2–124.9; P = 0.02) were preoperative factors associated with SSIs. Blood loss (OR = 1.0; 95% CI, 1.0–1.0; P = 0.039) and allografts and allografts in combination with other grafts were operative risk factors for SSIs. The final overall risk model for SSIs was weight-forage at the 95th percentile or higher (OR = 4.0; 95% CI, 1.4–⬁; P = 0.037), American Society of Anesthesiologists score 3 or more (OR = 3.8; 95% CI, 1.6–⬁; P = 0.01), and prolonged operation duration (OR = 1.0/min increase; 95% CI, 1.0–1.0; P = 0.004). SSIs were associated with 2.8 days of additional postoperative length of stay (P = 0.02). Neuromuscular scoliosis was the only factor significantly associated with hospital readmission (OR = 23.6; 95% CI, 3.8–147.3; P = 0.0007). From the *Department of Epidemiology and Public Health, University of Maryland School of Medicine, Baltimore; †Clinical Quality, Safety, and Process Improvement, University of Iowa Hospitals and Clinics, Iowa City; ‡Department of Epidemiology, University of Iowa College of Public Health, Iowa City; and Departments of §Internal Medicine, ¶Pediatrics, and Orthopedics, University of Iowa Carver College of Medicine, Iowa City. Acknowledgment date: March 19, 2014. First revision date: September 25, 2014. Second revision date: October 14, 2014. Acceptance date: October 29, 2014. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. Relevant financial activities outside the submitted work: travel/ accommodations/meeting expenses, board membership, grants. Address correspondence and reprint requests to Loreen A. Herwaldt, MD, FIDSA, Department of Internal Medicine, The University of Iowa Hospitals and Clinics, 200 Hawkins Dr, Iowa City, IA 52242; E-mail: loreen-herwaldt@ uiowa.edu DOI: 10.1097/BRS.0000000000000693
E112
www.spinejournal.com
Conclusion. Our results suggest that pediatric patients undergoing spinal fusion might benefit from antimicrobial prophylaxis that covers gram-negative organisms. Surgical duration, graft implantation, and blood loss are potentially modifiable operative risk factors. Neuromuscular scoliosis, high weight-for-age, and American Society of Anesthesiologists scores 3 or more may help surgical teams identify patients at high risk for SSI. Key words: surgical site infection, spinal fusion, pediatric, complication, scoliosis. Level of Evidence: 4 Spine 2015;40:E112–E119
S
urgical site infections (SSIs) after pediatric spine operations can have devastating consequences. A report from the Scoliosis Research Society indicates that the rate of SSIs for patients younger than 21 years undergoing spine operations is 1.4% for patients with idiopathic scoliosis, 2.2% for patients with congenital scoliosis, 5.5% for patients with neuromuscular scoliosis, and 5.7% for patients with post-traumatic scoliosis.1 Patients with myelomeningocele have the highest reported SSI rates after pediatric spine fusion procedures.2,3 Previous studies have identified possible risk factors both for SSIs1,4–12 and for major complications, including malnutrition,8 cognitive delay,10 absence of drains,7 blood loss,9 blood transfusions,7 cell count,11 use of allografts rather than autografts,10 a combined anterior and posterior surgical approach,12 and the underlying diagnosis.7,9 Most studies before 2000 were case series that lacked comparison groups, were too small for statistical analysis of risk factors,13–15 or included both adults and children and did not examine pediatric cases separately.6,9,16 Moreover, some studies included spinal procedures other than fusions17 and some assessed risk factors for all surgical complications.9,12,17–20 Some studies of pediatric populations evaluated only patients with specific operative indications, such as idiopathic scoliosis7,18,21,22 or neuromuscular scoliosis.8,11 Results of population-specific studies may not be generalizable to a broader population of children undergoing spinal fusions. The baseline SSI rate after pediatric spinal fusions at the University of Iowa Hospitals and Clinics was 2.59% during July 2001 to December 2006 and the rate increased to 3.92% during January 2007 to September 2009. Given increased SSI rates and gaps in our understanding of risk factors for SSIs in January 2015
Copyright © 2015 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE140399_LR E112
23/12/14 10:53 PM
SURGERY
Pediatric SSI Post–Spinal Fusion • Croft et al
these patients, we conducted a case-control study to identify modifiable risk factors.
RESULTS
MATERIALS AND METHODS
During the study period, 598 children (mean age = 13 yr) underwent 615 spinal fusions; 22 children acquired SSIs (3.6%). The mean time to onset of SSIs was 46 days (median = 26 d; range = 3–352 d). Eleven SSIs were superficial incisional and 11 were deep incisional infections. Nine of 11 patients requiring reoperations for SSIs had deep incisional infections. Wound samples from 16 case patients (72%) were cultured and all showed positive results. Of the 16 positive cultures, 10 (62.5%) cultures grew more than 1 organism, 12 (75%) grew at least 1 gram-negative organism, and 11 (69%) grew at least 1 gram-positive organism (Table 1). Five cultures grew Staphylococcus aureus; 1 isolate was resistant to methicillin. Four wound cultures grew coagulase-negative staphylococci, 3 grew Enterobacter cloacae, and 5 grew Pseudomonas aeruginosa. Six superficial SSI samples were not cultured.
We included pediatric spine procedures performed by 1 orthopedic surgeon at the University of Iowa Hospitals and Clinics during July 1, 2001, to July 31, 2010. We defined cases as patients who acquired infections meeting the Centers for Disease Control and Prevention’s National Health Safety Network’s definition of SSI.23 Twenty-two of 598 pediatric patients undergoing spinal fusions during the study period acquired SSIs. We matched each case with 2 patients without SSIs (controls): the patient who underwent spinal fusion immediately before the case and the patient who underwent spinal fusion immediately after the case. We abstracted demographic data (e.g., age, sex, and insurance status) and data on numerous variables from the patients’ medical records: preoperative weight, height, body mass index, underlying diseases, medications, prior spine operations, preoperative laboratory values (hemoglobin, lymphocyte count, glucose [minimum and maximum within 24 hr before the operation and 48 hr after the operation], hemoglobin A1c, albumin, urine analysis, and urine culture), length of hospital stay, the American Society of Anesthesiologists (ASA) score, surgical indication, surgical approach, instrumentation type, bone graft type, blood loss, transfusions, drain placement, lowest intraoperative temperature, lowest intraoperative fraction of inspired oxygen, procedure duration, wound culture results, antimicrobial therapy, seroma formation, hematoma formation, and reoperations. Data were entered into a Microsoft Excel spreadsheet and analyzed using SAS 9.2 (SAS Institute, Cary, NC). We used the Fisher exact test to analyze categorical variables and the Student t test to analyze continuous variables. If data on comorbidities, preoperative medication, antimicrobial prophylaxis, or transfusion were not available, we assumed that the patient did not have the comorbidities or did not receive the treatment. We constructed preoperative, operative, and overall (included patient-related and procedure-related factors) multivariable models to identify factors associated with SSIs. We included variables without missing values and with P ≤ 0.10 or that we presumed to have biological significance in multivariable ordinary logistic regression models. We used stepwise selection to construct an overall multivariable model; P ≤ 0.15 was required for both entering and remaining in the final model. After we identified the final models, we used conditional logistic regression to adjust for the clustering effects of matched case patients and controls. We used linear regression to identify variables that influenced postoperative length of stay and an ordinary logistic regression model to identify variables predicting readmission. We assessed the predictive power of each ordinary logistic regression model with the c-statistic, where 0.5 and 1.0 correspond to no predictive power and perfect predictive power, respectively. We used the Hosmer-Lemeshow goodness-of-fit test to assess how well the models fit the data. Variables with P values of less than 0.05 were considered statistically significant. Spine
Surgical Site Infections
Bivariable Analyses Case patients and controls did not differ significantly by age, height, weight, most comorbidities, and numerous surgical
TABLE 1. Combinations of Organisms Cultured
From Surgical Site Infections
Organism
No. Case Patients (N = 22)
Bacteroides fragilis, Enterobacter cloacae, Staphylococcus aureus
1
B. fragilis, mixed anaerobes including Proteus vulgaris
1
CNS
2
CNS, Haemophilus parainfluenzae, α-hemolytic streptococci
1
E. cloacae
1
Escherichia coli, Pseudomonas aeruginosa
1
Enterococcus spp., Proteus mirabilis
1
Enterococcus spp., E. coli, Klebsiella pneumoniae, CNS
1
Enterococcus spp., K. pneumoniae, P. aeruginosa (2 strains)
1
MRSA, Propionibacterium acnes
1
P. aeruginosa
1
P. aeruginosa (2 strains)
1
S. aureus
1
S. aureus, E. cloacae
1
S. aureus, P. aeruginosa (2 strains)
1
No culture
6
CNS indicates coagulase-negative staphylococci; MRSA, methicillin-resistant Staphylococcus aureus.
www.spinejournal.com
E113
Copyright © 2015 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE140399_LR E113
23/12/14 10:53 PM
SURGERY
Pediatric SSI Post–Spinal Fusion • Croft et al
TABLE 2. Bivariable Analysis* Patients With SSIs (N = 22)
Control Patients (N = 44)
Odds Ratio (95% CI)
P
13.1 ± 3.7
13.4 ± 3.0
…
0.73
Male sex
11 (50.0)
16 (36.4)
1.8 (0.6–4.9)
0.30
No private insurance
3 (13.6)
11 (25.0)
0.5 (0.1–1.9)
0.35
144.4 ± 25.3
153.9 ± 19.1
…
0.09
53.2 ± 26.4
49.0 ± 21.8
…
0.49
8 (36.4)
2 (4.6)
12.0 (2.3–63.3)
0.002
23.9 ± 5.9
19.1 ± 2.9
…
