Abstract
Abstract
Background:
Surgical site infection (SSI) remains a major clinical problem for developing countries in terms of morbidity, mortality, and hospital cost. Little is known about the epidemiology of SSI in Nepal. We conducted a study in Nepal to identify the various pre- and intra-operative risk factors for SSIs that are accessible to interventions.
Methods:
The study was a prospective study done on all eligible patients who underwent abdominal surgery in the surgical wards of Tribhuvan University Teaching Hospital (TUTH) in Kathmandu, Nepal, from January 2011 to June 2011. We evaluated some patient-related as well as intra-operative variables that could be risk factors for SSIs. We assessed the association between these risk factors and SSI with the Fisher exact test and logistic regression analysis.
Results:
Of the 230 patients included in the study, 53 were identified as having a SSI, resulting in an overall rate of SSI of 23%. Multivariable analysis identified the following factors as independent risk factors for SSI: (1) Low hemoglobin concentrations (<12 g/dL) (odds ratio [OR] 2.5; 95% confidence interval [CI] 1.1–6.1); (2) overweight (OR 7.6; 95% CI 2.1–27.0); and (3) surgery performed by residents (OR 3.4; 95% CI 1.4–8.3).
Conclusions:
Surgical site infection is common among patients undergoing abdominal surgery at TUTH. This study identified some preventable risk factors associated with SSI at TUTH. Identification of such risk factors is expected to help surgeons improve patient care and decrease mortality and morbidity as well as the hospital-care cost of surgical patients.
Post-operative SSIs are common complications in patients undergoing gastrointestinal surgery, occurring in 10%–30% of such patients [3]. The severity of SSI ranges from mild, requiring local wound care and antibiotics, to serious, requiring re-operation and incurring high mortality rates. These complications prolong hospital stays and result in an increased cost of health care [4,5].
The National Nosocomial Infection Surveillance study identified wound class, American Society of Anesthesiologists (ASA) Physical Status Classification, and prolonged operative time as risk factors for SSI [6]. However, increasing emphasis has been given more recently, to systemic factors such as age, gender, lifestyle, and coexisting morbidities, which are considered determinants in the pathogenesis of SSIs [7,8].
Prospective studies of the incidence and risk factors for SSI in Nepalese hospitals are generally lacking. A few retrospective studies in Nepal have suggested the prevalence rate of SSI to be 4%–7% for all kinds of operation [9,10]. A better understanding of the risk factors associated with such infections could help reduce their occurrence by promoting effective strategies for prevention. The study described here documents multiple factors predicting SSI in a tertiary-care hospital in Nepal.
Patients and Methods
Study design
The protocol for this prospective study was reviewed by the Institutional Review Board of the Institute of Medicine of Tribhuvan University. The study was done in the Department of Surgery of Tribhuvan University Teaching Hospital, a 444-bed tertiary-care hospital in Nepal. All patients undergoing open gastrointestinal surgery in the Department of Surgery from December to June 2011 were assigned to clinical follow-up and collection of their data.
Study participants
Subjects included in the study had undergone open gastrointestinal surgery or surgery involving opening of the peritoneum. Patients who underwent esophageal, anal/perianal, or laparoscopic surgery of any kind were excluded from the study. Patients were informed about the study both orally and in writing. Participation was voluntary, and patients had the right to stop their involvement in the study without being asked why.
Data collection
After the acquisition of patient consent for participation in the study, necessary information was gathered through patient interviews and from hospital data when necessary. The risk factors for SSI included in the assessment were age, gender, co-morbidities, use of immunosuppressive medications, body mass index, smoking, alcohol use, hemoglobin concentration, lymphocyte count, duration of surgery, type of surgery (emergency/elective), qualifications of the surgeon (attending vs. resident), changing of surgical gloves during surgery, and classification of surgical incision (clean, clean-contaminated, contaminated, or dirty/infected).
The outcome measure for the study was SSI. The criteria developed by the CDC and the National Nosocomial Infections Surveillance System was used for the diagnosis of SSI [1].
After surgery, patients were assessed daily during hospitalization. Clinical examination of the incision site was done on a daily basis by a consultant surgeon. In all cases, dressing changes were made daily by a surgical resident during the patient's hospitalization. In case of a suspected SSI, swab specimens from the incision site were taken and incubated overnight. All swab specimens were analyzed by gram staining and culture on agar in the hospital laboratory. Routine post-operative care was provided to each patient, and follow-up was done for a minimum of 30 d to detect SSIs that might have developed after a patient's discharge.
Data analysis
Data analysis was done with SPSS version 17.0 software (SPSS, Chicago, IL). Complete confidentiality was maintained while the data were being processed. Descriptive statistics of all variables were computed and used as the primary method of data analysis. Differences in the distribution of clinical data and in the development of SSIs were evaluated with the Fisher exact test for categorical variables. Multiple logistic regression analysis was used to examine the predictors of SSI. Following univariate analysis, variables that were statistically significant were used in a multivariable analysis. All results were described with odds ratios (ORs) and 95% confidence intervals (CIs). The level of significance chosen was 5%.
Results
A total of 230 eligible patients were observed during the study period, of whom 132 (57.4%) were male and 98 (42.6%) were female. Seven patients (3%) were excluded from the analysis because of a lack of follow-up. The mean age of the patients was 40 years, with a standard deviation of 17.4 years. A total of 53 patients developed SSI, thus giving an overall infection rate of 23.04%. Among the 53 cases of SSI, 51 (96%) had superficial SSIs and two (4%) had organ/space infections. Table 1 presents additional characteristics of the study patient population.
Co-morbidities include diabetes mellitus, chronic kidney disease, chronic obstructive pulmonary disease, chronic liver disease, congestive heart failure, and human immunodeficiency virus infection/acquired immune deficiency syndrome and other immunocompromised states.
Others include abscess drainage, hernioplasty, stoma closure, intestinal arterio-venous malformation, and splenectomy.
SSI=surgical site infection.
Pre-operative factors
Patient factors found to be significantly associated with the development of SSI were a history of smoking (p=0.002), alcohol consumption (p=0.01), co-morbidities (p<0.001), low hemoglobin concentration (p<0.001), use of immunosuppressive medications (p<0.001), increasing age (p=0.01), and overweight (p<0.001). There was no statistically significant association between absolute lymphocyte count (p=0.28) or gender (p=0.25) and the development of SSI. Details of the bivariate analyses of the various pre-operative factors with SSI are shown in Table 2.
CI = confidence interval; OR = odds ratio; SSI = surgical site infection.
Intra-operative factors
Factors significantly associated with the development of SSI were changing of gloves during surgery (p<0.001), surgery performed by a resident (p<0.001), and duration of surgery >3 h (p<0.001). We did not see any significant differences among incision types and the development of SSI (p=0.65). None of five cases of clean incisions was complicated by an SSI. The infection rates for the remaining types of incisions were 22.6% for clean-contaminated, 15.5% for contaminated, and 38.3% for dirty. In terms of complication by an SSI, emergency operations were no different than elective ones (p=0.41). Table 3 summarizes the association among various intra-operative factors and the development of SSI.
CI = confidence interval; GI=gastrointestinal; OR = odds ratio; SSI = surgical site infection.
Multivariable analysis
On the basis of the respective bivariate analyses of the various pre-operative and intra-operative risk factors and SSI, as shown in Tables 2 and 3, variables that showed a statistically significant (i.e., p<0.05) association with SSI were selected for a multivariable analysis. A total of 10 different variables were subjected to the multivariable logistic-regression model, using the “Enter” method. Table 4 shows the details of the overall model including the ORs and CIs of the different predictor variables. The discrimination function of the model was tested through a receiver operating characteristic curve analysis of the predicted probabilities, which revealed an area under the curve of 0.8. Calibration data for the model were calculated with the Hosmer–Lemeshow test, which revealed a χ2 value of 4.8 with a value of p=0.778.
BMI = body mass index.
On multivariate analysis, the following factors were significantly found to predict SSI: a low hemoglobin concentration (<12 g/dL), overweight, duration of surgery>3 h, and having surgery done by a resident, as shown in Table 4.
Bacterial infection and susceptibility pattern
Of the 53 patients with a diagnosis of SSI and whose specimens were collected for microbiologic investigation, 47 (88.6%) had positive bacterial growth on culture within 48 h of incubation. Six of these patients' 47 cultured specimens (8.9%) had mixed growth. The most common organism isolated was E. coli, in 23 (43%) of the specimens, followed by Proteus in 4 (7.5%) specimens, Acinetobacter in 4 (7.5%), Staphylococcus aureus in 3 (5.6%), and Pseudomonas in 2 (3.7%), as shown in Table 5. Two of the three S. aureus isolates were methicillin-resistant but vancomycin-sensitive. Of the 23 E. coli specimens isolated, 16 (69.5%) were resistant to ciprofloxacin and 12 (52.1%) were sensitive to amikacin.
Others include Peptostreptococcus, Klebsiella, and Prevotella.
Post-operative care and use of antibiotics
All patients received routine post-operative care including antibiotic prophylaxis. The most frequent antibiotic prescribed was ceftriaxone (185 cases [80.4%]), followed by cefixime (26 cases [11.3%]) and ciprofloxacin (10 cases [4.3%]). Anaerobic coverage, consisting predominantly of metronidazole, was added in 215 cases (93.5%).
Discussion
Post-operative SSI remains one of the most important causes of morbidity in patients treated surgically. These infections incur higher cost because of longer hospitalizations, more nursing care, additional wound care, potential readmission to the hospital, and further surgical procedures [11]. Intensive infection surveillance and control programs can reduce the rates of infection by as much as 35%–50% [12,13].
The overall infection rate in this study was 23%. This was remarkably higher than the infection rates reported in developed countries such as Italy (5.9%), the United States (4.3%), and Denmark (6%) [11,14,15]. However, studies conducted at hospitals in developing countries indicate similar rates of infection among them, such as in Tanzania (23%), Ethiopia (21%), and Peru (26%) [16–18]. These latter findings are unsurprising, given that no surveillance and feedback program for SSI has been implemented in these countries, in what is a scenario common in developing countries. It also reflects a lack of adequate post-operative care and failure to maintain sterility during surgical procedures. Prevalence studies of SSI previously conducted in Nepal showed much lower infection rates [9,10]. However, these studies were retrospective and did not investigate SSIs that developed after discharge from the hospital, thereby underestimating the rate of infection. Moreover, these studies examined SSIs for all types of operations in contrast to the present study, which considered only abdominal operations. Abdominal operations tend to have high infection rates as compared to other kinds of operations, thus explaining our high infection rate [19,20].
Our patients were relatively young, with a mean age of 40 years. This is unsurprising given that the two major diagnoses in our study, appendicitis and benign biliary tract disease, are most common in young and middle-aged adults. Increasing age, current smoking, alcohol consumption, the presence of co-morbidities, a low hemoglobin concentration, the use of immunosuppressive medications, and overweight were the patient factors significantly associated with SSI in the bivariate analysis done in our study. These findings are concordant with those in the existing literature. Smoking has been shown to be an independent risk factor for SSI [15,21–23]. Smoking delays the healing of SSIs by causing local and systemic vasoconstriction. This results in tissue hypoxia and hypovolemia, an environment conducive to SSI [22]. Heavy alcohol consumption weakens immunity and increases the risk of SSI [14], although this effect is dose-dependent and was not addressed in our study.
Increasing age was associated with a higher incidence of SSI. This finding was in agreement with that in previous studies [11,21,23–25]. Overweight is an independent risk factor for SSI, as has been confirmed by previous studies [11,21–23,26–28]. In the OR, an obese patient presents technical challenges that often increase procedure time and complications [26]. Because adipose tissue is vascularized poorly, it creates an environment suitable for the proliferation of microorganisms. Previous studies have documented that co-morbidities are associated with an increased risk of SSI [14,15,21,26]. This is in agreement with our finding. Similarly, a low hemoglobin concentration creates the risk of SSI through tissue hypoxia, as has been proved in the past [14].
Changing of gloves during surgery, the performance of surgery by residents, and a prolonged duration of surgery were the intra-operative factors significantly associated with SSI in our bivariate analysis. Frequent glove changing may have been associated with a higher incidence of SSI because it often takes place in procedures in which there is more contamination, such as colorectal surgery and operations involving a stoma. Previous studies have documented this association [22].
Previous studies have also assessed the influence of prolonged operative time as a risk factor for SSI [16,21,23,26,27,29]. A prolonged operative time leads to fatigue, resulting in a decline in the use of aseptic measures during surgery, and may also be associated with advanced disease, re-operation, or intra-operative difficulties. Additionally, a prolonged operative time is often related to increased blood loss, which contributes to tissue hypoxia [22].
A further observation resulting from our study was that a higher seniority of surgeons was significantly associated with a reduced rate of SSI. The risk of infection was on average five-fold higher in operations performed by residents than in those performed by attending surgeons. This could be due to inadequate surgical skills, longer durations of surgery, and greater blood loss during surgery done by residents. As with some of the other findings in our study, previous studies have documented this finding [30–32].
Interestingly, wound classification was not an independent predictor of SSI in our study. A possible reason for this was that most of the pathogens isolated in the study were exogenous rather than endogenous. These pathogens may have been introduced by the penetration of contaminated instruments into deeper tissue layers during surgical intervention. Our hospital has a high frequency of surgical interventions, with limited resources. Limited time and manpower can jeopardize the adequate preparation and sterilization of surgical instruments for re-use.
As documented in previous studies, we found no difference in the rates of SSI in observed elective and emergency operations [18,26]. Nor did we find a difference in the rates of SSI in male and female patients, which is also consistent with the findings in previous studies [11,18].
Among the 53 cases of SSI in our study, most (96%) were of superficial incisional type, with only two cases of SSI (4%) involving an organ space. Studies have shown that SSI involving an organ space is the least common type of SSI, representing from 2% to 5% of such infections [11,25].
The most common organism isolated in our study was E. coli, which was in agreement with findings reported previously [17,33,34]. Resistance to ciprofloxacin among the organisms isolated in our study was more common than expected. This underscores the importance of understanding the local sensitivity patterns of microorganisms isolated from SSIs for ensuring their effective treatment.
By logistic regression analysis, we found that a low hemoglobin concentration (<12 g/dL), overweight, a prolonged duration of surgery, and surgery performed by a resident were significant predictors of SSI in our patient population. Fortunately, all of these risk factors are preventable. Adequate pre-operative preparation, reducing the duration of surgery and adequate training in proper surgical techniques, improving the skills of junior surgeons, and direct supervision of trainees by more experienced surgeons can help reduce the rates of SSI at our hospital. Because infection surveillance programs have been shown to reduce rates of infection remarkably, we see the need for such a program in our hospital [12,13,23,35].
Our study had several limitations. The size of the cohort was relatively small, particularly in terms of the subsets of patients with deep incisional and organ/space SSIs. Culture from swab specimens, which was often used in our study for the diagnosis of SSI, has its limitations. Cultures obtained from swab specimens reflect only the bacteria colonizing the surface of an anatomic site rather than pathogenic strains involved in tissue invasion [36]. Swab cultures can also miss anaerobic and some fastidious bacteria [36], and have limited sensitivity and specificity in identifying SSI, which depends on the swabbing technique [37]. Similarly, our study collected no data on other measures shown to affect the incidence of SSI, such as ASA Physical Status Classification, the method and timing of hair removal at a surgical site, skin preparation for surgery, the adequacy of mechanical bowel preparation, glucose management, hypothermia, and the duration of a patient's pre-operative hospital stay. Additionally, the study did not include an analysis of the dose-related effect of alcohol consumption on the incidence of SSI.
Acknowledgment
We thank Tejaswi Giri and Sudha Regmi for their help in editing the manuscript of this paper.
Footnotes
Author Disclosure Statement
No competing financial interests exist.