Surgical Site Infection Rates and Risk Factors in Orthopedic Pediatric Patients in Madrid, Spain

Abstract

Background: 

The objective of the study is to study surgical site infection (SSI) rates and risk factors in a pediatric population.

Methods: 

We conducted a prospective cohort study to estimate the SSI rate at a national pediatric referral center, covering all patients managed at the Orthopedic Surgery Department of the Niño Jesús Children’s University Teaching Hospital from January 2010 through December 2012. Risk factors and antibiotic prophylaxis were monitored. A comparison between Spanish and US data was performed, with a breakdown by National Nosocomial Infection Surveillance risk indices. We also conducted a comparative study of SSI rates from 2010 to 2012 to assess the impact of the epidemiologic surveillance system.

Results: 

The study population of 1079 patients had a SSI rate of 2.8%. SSI rates were calculated for spinal fusion and other musculoskeletal procedures according to the National Nosocomial Infection Surveillance risk index. In the case of other musculoskeletal procedures, our SSI rates were 0.8 times lower than the overall Spanish rate, but higher than US rates for all risk categories. For spinal fusion procedures, our SSI rates were 1.2 times higher than the Spanish rates and 3.5 times higher than National Nosocomial Infection Surveillance rates. This latter finding should be interpreted with caution because it was based on a small sample. The multivariate analysis indicated that the only predictive factors of SSI were American Society of Anesthesiologists score and age.

Conclusions: 

The surveillance program showed that for clean procedures, SSI incidence decreased from 4% in 2010 to 3.2% in 2011 and to 2.4% in 2012.

Healthcare-associated infections (HAIs) are among the leading causes of morbidity and mortality in hospitalized patients.^1,2 Surgical site infections (SSIs) are the second most common type of HAIs and account for approximately 20% of all infections.³ According to a National Healthcare Safety Network (NHSN) report, the overall surgical site infection (SSI) risk from January 2006 to December 2008 was 1.9 per 100 procedures (range 0.26–13.83).⁴ This may be partially explained by the increasing number of surgical patients who have a wide variety of underlying chronic diseases. In addition, more prosthetic implant and organ transplant operations are being performed.⁵ Recent evidence suggests that HAIs can be prevented through implementation of evidence-based best practices.⁶

In contrast to the large body of adult literature on this topic, there is a shortage of data on SSIs among children. Antibiotic protocols for the prevention of infections issued by the Centers for Disease Control and Prevention do not contain sections designed for children. Hence, when it comes to developing prevention guidelines, infections in children are still insufficiently understood.⁷

Comparison of SSI rates over several years or among different groupings can help assess a new surgical technique or prophylaxis. To avoid bias, however, any such assessment must take into account the medical condition of the persons included in the comparison groups.⁸

The 3 main factors that influence the likelihood of SSI are: the degree of contamination during the operation, the patient’s general health and the duration of the surgical intervention. These are the surveillance parameters used for the National Nosocomial Infection Surveillance (NNIS) index, which classifies patients according to their potential risk of infection.⁹ Despite being useful for modeling SSI risk in adults, this prediction model may not be applicable to children, due to significant differences in age, underlying medical conditions, procedures and care processes. To date, there is limited evidence defining risk factors associated with development of SSI in a pediatric population.¹⁰

Accordingly, the aims of this study were as follows: to estimate SSI incidence in a pediatric orthopedic surgery department and classify it according to the NNIS index; to assess the extent of compliance with the antibiotic prophylaxis protocol; to identify possible SSI-related risk factors and to assess the impact of an epidemiologic surveillance system on SSI rates across the study period.

PATIENTS AND METHODS

We conducted a prospective cohort study from January 2010 through December 2012 to determine the SSI incidence rate at the Pediatric Orthopedic Surgery Department of the Niño Jesús Children’s University Teaching Hospital in Madrid, which is a national referral center for high-quality pediatric care.

All patients attended at the above orthopedic surgery department across the study period were included in the analysis. The Preventive Medicine Department undertook data collection and clinical monitoring from the date of admission to the date of discharge. Patients were not monitored after discharge owing to a lack of resources. The criteria used to define SSI and patient risk index categories were those established by the Centers for Disease Control and Prevention and NNIS.

All data were recorded using a monitoring application developed by the Spanish Program for Clinical Indicators of Continuous Quality Improvement [Indicadores Clínicos de Mejora Contínua de Calidad (INCLIMECC)]. Our program classifies antimicrobial prophylaxis as adequate or inadequate according to our hospital’s antibiotic prophylaxis protocol. In addition, inadequate prophylaxis is further classified as: inadequate choice (where the agent used differs from that recommended in the protocol, with no reason given, eg, allergy); inadequate time of initiation (where the agent is initiated: >5 minutes before induction of anesthesia, except in the case of vancomycin or after surgery); inadequate application method (where antibiotic prophylaxis is administered despite not being indicated, as in clean surgeries without prosthetic implants) or inadequate duration (where antimicrobial prophylaxis is continued for more than 24 hours). Our antibiotic prophylaxis protocol is shown in Table 1. The variables analyzed were sex, age, patient characteristics, surgical procedure codes, elective or emergency surgery, laparoscopic versus open surgery, presence of implants, duration of the intervention, surgical risk according to American Society of Anesthesiologists (ASA) score, degree of contamination of surgery, administration of prophylactic antibiotic and adequacy of prophylaxis (choice, time of initiation, application method, duration). Procedures were classified according to International Classification of Diseases, Ninth Revision, Clinical Modification categories and SSI definition criteria, and sites were named in line with Centers for Disease Control and Prevention definitions.

TABLE 1: Antibiotic Prophylaxis Protocol in Orthopedic Surgery

A descriptive study was performed to establish patients’ sociodemographic characteristics. Qualitative variables were described with their frequency distribution, and compared using Pearson’s χ² test or, where the conditions for this were not present, using Fisher’s nonparametric exact test. Quantitative variables were analyzed using the Student’s t-test where the figures followed a normal data distribution or the Mann-Whitney U test where they did not. For quantitative variables in >2 groups, we used analysis of variance or, alternatively, the nonparametric Kruskal-Wallis method in cases where test application conditions were not met. A multivariate logistic regression analysis was used to find independently associated risk factors for SSI, with these risk factors being automatically selected using the backward stepwise method. Procedures were divided into 2 categories, namely, “spinal fusion” and “other musculoskeletal”.

Infection rates, both overall and stratified according to the NNIS index, were calculated for each procedure. Similarly, the average lengths of preoperative, postoperative and total hospital stays were also calculated.

To assess the impact of the epidemiologic surveillance system throughout the study period, we compared the overall and stratified SSI rates, microorganisms implicated, adequacy of surgical prophylaxis and lengths of hospital stay across the study period.

A comparison between Spanish and US data was performed, with a breakdown by NNIS risk indices. For each procedure, we computed both the observed rate and the predicted rate based on the latest figures published by the NHSN and the INCLIMECC network in Spain, using the indirect standardization method. The standardized infection ratio is the observed number of infections divided by the expected number of infections and is interpreted as the relative risk.¹¹

RESULTS

The study covered a total of 1079 patients, mean age 8.9 years [range: 0 (6.7%) to 20 (0.4%)]. Patient characteristics, diagnosis and surgical procedures are shown in Table 2. Mean surgical duration was 86.1 minutes (standard deviation 82.2), with duration of surgery being above the 75th percentile in 14 of the 30 SSI cases and in 291 of the 1049 “non-SSI” cases. Antibiotic prophylaxis was administered to 55.1% of all patients. Overall adequacy of the antibiotic prophylaxis protocol was 54.3%. The causes of inadequate antibiotic prophylaxis were: application method (42.3%), duration (40.1%), time of initiation (12.9%) and choice (0.7%). There was also a small percentage (4.2%) of patients undergoing surgery, who should have received but were never given a dose of a prophylactic antibiotic.

TABLE 2: Patient Characteristics and Diagnostic and Surgical Procedures (%)

The average preoperative hospital stay was 1 day. The average total length of stay for patients undergoing orthopedic surgery was 3.6 days, with this rising to 8.8 days in the case of those with SSIs. The average postoperative hospital stay was 2.5 days, increasing to 7.8 days for patients with SSIs.

Cumulative SSI incidence was 2.8%. A breakdown by degree of surgical contamination showed that clean sites had an SSI rate of 2.9% and dirty sites a rate of 1.2%. The infection rate classified by depth was as follows: 1.4% superficial incisional; 1% deep incisional and 0.4% organ/space. SSI incidence increased in parallel with the NNIS index for both procedures.

In descending order of frequency, the individual microorganisms implicated were Staphylococcus aureus (28.6%), Staphylococcus epidermidis (9.5%), Enterobacter cloacae (7.1%) and Methicillin-resistant S. aureus (7.1%). The first-line antibiotic selected for SSI treatment was amoxicillin-clavulanate.

The univariate analysis showed that the factors associated with SSI were age ≥14 years, ASA score ≥3 and duration of surgery higher than the 75th percentile. The multivariate analysis showed that the only predictive factors were ASA score and age, with the other factors proving nonsignificant.

For spinal fusion procedures, the SSI rate in orthopedic surgery at our health center was 1.2 times higher than that expected on the basis of the INCLIMECC figure and 3.5 times higher than that expected on the basis of NHSN data. For other musculoskeletal procedures, the SSI rate in orthopedic surgery was: 0.8 times lower than the INCLIMECC figure and, on the basis of NHSN data, 3.1 times higher than that expected for risk index category 0, 5.2 times higher than that expected for risk index category 1 and 1.5 times higher than that expected for risk index categories 2 and 3.

One of the stated aims of our study was to observe how the implementation of a surveillance program might serve to reduce incidence of SSI in a pediatric orthopedic surgery department. This monitoring system showed that, in clean procedures, SSI incidence decreased from 4% in 2010 to 3.2% in 2011, followed by a further decrease to 2.4% in 2012; and that in dirty surgery, SSI was solely in evidence in 2011, with a rate of 2.9%.

Comparison of SSI rates by depth showed that, while rates for superficial and deep wounds decreased, those for organ-space wounds remained fairly constant throughout the study period. In terms of the germs involved in SSIs, S. aureus was the most commonly isolated bacteria over the 3 years of study.

Examination of the trend in SSI rates according to the NNIS risk index showed: in the case of spinal fusion procedures, a decline throughout the study period in every risk index; and in the case of other musculoskeletal procedures, a decrease for risk index 1 and an increase from 2010 to 2011 which was maintained in 2012 for risk indices 2 and 3, with a decline in the overall rate across the entire study period.

The trend in adequacy of surgical prophylaxis was quite favorable, increasing from an adequate 44.7% in 2010 to 57.1% in 2011, and slightly further still to 59.8% in 2012. The most frequent causes of inadequacy of prophylaxis were indication of prophylaxis in 2010 and 2012 and length of prophylaxis in 2011.

DISCUSSION

Few studies have assessed SSI incidence and risk factors in children. Many literature addressing SSI risk has been generated from the study of adult surgical patients. Casanova et al⁸ prospectively studied 3700 children admitted postoperatively to a single university hospital. They identified 8 factors associated with development of an SSI, that is, wound classification, type of operation, duration of operation, use of peripheral venous catheter, use of central venous catheter, use of urinary catheter, number of diagnoses and postoperative length of stay. Other studies have identified similar factors in patients undergoing cardiothoracic, orthopedic and neurosurgical procedures.^12–14 SSI rates vary widely among studies, and comparisons are rendered difficult by differences in patient selection, definitions of infection surveillance methods and statistical analysis of risk factors. Duque-Estrada et al¹⁵ conducted a 7-year retrospective study covering a total of 537 patients who underwent 575 operations and reported SSI rates of 2.7% for clean wounds, 10.5% for clean-contaminated wounds, 13.5% for contaminated wounds and 14.6% for dirty wounds. A prospective study conducted by Varik et al¹⁶ on SSI incidence in children over a period of 2 years showed an overall SSI rate of 1.2%, which broke down as 0.5% for clean, 1% for clean contaminated, 5.3% for contaminated and 13.3% for dirty surgeries. For orthopedic operations, the SSI rate was 1.4%.

The main goals of HAI surveillance are to ascertain rates and trends, detect changes in patterns and identify the microorganisms and risk factors involved, with a view to designing measures aimed at reducing incidence. Surveillance standardizes information, thereby enabling comparison with data from other health centers. The direct method, with daily observation of the surgical site by specialist medical staff trained in infection prevention and professional follow up starting 24–48 hours after surgery, is the most accurate method of surveillance.¹⁷

According to 1 study on pediatric orthopedic surgery,¹⁸ age <2 years is a risk factor for development of SSI. Both the univariate and multivariate analyses in our study indicated, however, that age 14 years or over was a risk factor for development of SSI. This result is not consistent with earlier studies and might be attributable to the following fact: SSIs were most frequently seen in medical conditions such as scoliosis, recurrent knee dislocation and supracondylar fracture of the humerus, the first 2 of which are more frequently found in older children. Scoliosis surgery is complicated and lasts longer than other procedures in orthopedic surgery. Similarly, the univariate and multivariate analyses also showed that an ASA score of more than 2 was a risk factor for development of SSI, a finding in line with that of Linam et al¹⁴ in a study on SSI incidence in spinal fusion.

Adherence to antibiotic prophylaxis protocols was low. Comparison with other health centers nationwide showed our results to be worse than those reported for other hospitals in Spain, namely: inadequate prophylaxis, 45.7% versus 27.5% for the Madrid Region (Comunidad de Madrid)¹⁹ and 27.7% for Spain as a whole; not receiving prophylaxis despite its being indicated, 4.2% versus 2.3% for the Madrid Region and 1.9% for Spain as a whole, bearing in mind that, for the purposes of this study, these data include all surgeries performed. Among patients who developed SSIs, average hospital stay was 8.8 days at our health center, 41 days in the Madrid Region and 31.4 days for Spain as a whole.

The average length of stay at our health center was below that of other centers in the country. This is probably because the hospitals used for comparison purposes are adult facilities, whereas ours is a pediatric center to which children are normally admitted on the day of intervention or the previous day. Moreover, orthopedic department patients do not usually have complex underlying pathologies which, if added to an SSI, would then complicate their hospitalization. SSI rates at our health center (2.8%) were below both the national (5.9% for the Madrid Region and 4.2% for Spain) and US rates (5.6%).

SSI rates in children display a wide variation among studies, with reported rates ranging from 0.9% (Bucher et al)¹⁰ through 18.7% (Porras-Hernandez et al).²⁰ Indeed, our rates are lower in comparison with those observed by other authors who have undertaken studies on orthopedic surgery in children (Kaabachi, 8.3%;¹⁸ Milstone, 3.4%;²¹ Linam, 4.4%;¹⁴ Ballard²² 4.5%). Three of these authors (Milstone, Linam and Ballard) conducted their studies on the spinal fusion procedure, in which surgery lasts longer than it does in other orthopedic procedures.

SSI rates were calculated for spinal fusion and other musculoskeletal procedures according to the NNIS risk index. In the case of other musculoskeletal procedures, our SSI rates were lower than those of the INCLIMECC network but higher than US rates for all risk categories. In the case of spinal fusion procedures, our SSI rates were higher than both the INCLIMECC and NNIS rates but this finding should be interpreted with caution since it was based on a small sample. Analysis of the expected rate using indirect standardization and the standardized infection ratio showed results that were similar to or better than those yielded by other Spanish hospitals but worse than those reported for US hospitals.

A prospective epidemiologic SSI surveillance system is in place at our health center. Cumulative SSI incidence described a clear downward trend across the study period. This finding is consistent with those of other authors, who have also observed a decrease in SSI rates in response to the implementation of a surveillance system.²³ Studies conducted by Cruse et al. and Olson show that, over a 10-year period, SSI rates decreased from 2.5% to 0.6%²⁴ and from 4.2% to 1.8%,²⁵ respectively

Adequacy of surgical prophylaxis showed a fairly favorable trend. This finding highlights the importance of feedback being channeled to participating departments, so that measures can be designed to reduce SSI incidence. At all events, the current degree of compliance with the prophylaxis protocol still shows room for improvement.

One of this study’s limitations was the absence of active postdischarge surveillance, bearing in mind that many SSIs are only manifested after the patient has been discharged.^26–28

The risk factors identified by our study were ASA scale score >2 and age ≥14 years. In the univariate analysis, duration of surgery ≥75th percentile was also statistically significantly associated with risk of developing an SSI but this association was not confirmed by the multivariate analysis. Other studies on children have concluded that operation-related factors play a more important role in determining risk of SSI than does the patient’s general physiologic condition.

The only intrinsic factors analyzed were sex, age and global physical status as indicated by ASA score. Other factors that are not included in the current monitoring system but may nonetheless be involved in SSI are diabetes, use of immunosuppressants, presence of neoplasms, use of drains, perioperative hypothermia and hypoxia. This could be a seen as a shortcoming of our study.

In conclusion, the SSI rate at our health center during the study period was lower than that of other hospitals in Spain which also participate in the INCLIMECC network, and lower than the US rates published by the NHSN. Indirect standardization and the standardized infection ratio show our rates to be similar to or lower than the overall Spanish rate but higher than US rates.

Adult-based groups were used for comparison purposes, in view of the fact that comparisons with other pediatric-based studies were rendered difficult by methodologic differences. Although SSI rates were calculated according to the NNIS risk index, it has not been yet validated for the pediatric population. Implementation of an epidemiologic surveillance system with feedback to the orthopedic department has succeeded in bringing about a gradual reduction in the SSI rate and a slight improvement in the degree of adequacy of surgical prophylaxis.