Identification of Risk Factors for Post Surgical Wound Infections in Elective Operations: A Multivariate Statistical Analysis

Dr. Md. Abdul Mazed Chowdhury1 and Dr. Aivee Ferdous2 1 Associate Professor in Statistics, Department of Accounting and Information Systems, University of Rajshahi, Rajshahi-6205, Bangladesh 2 Registrar, Enam Medical College Hospital, Savar, Dhaka, Bangladesh

ABSTRACT Surgical site infections are the most common complications of inpatient admissions and have serious consequences for outcomes and costs. The significant risk factors or variables, which affect the abdominal surgical site infections and their incidence are: age, sex, nutrition and immunity, prophylactic antibiotics, operation type and duration, type of shaving and secondary infections.

INTRODUCTION

The ancient Egyptians were the first civilization to have trained physicians to treat physical ailments. Medical papyri, such as the Edwin Smith papyrus (circa 1600 BC) and the Ebers papyrus (circa 1534 BC), provided detailed information of management of disease, including wound management with the application of various potions and grease to assist healing (Breasted, 1930; Bryan, 1930).
Hippocrates (Greek physician and surgeon, 460-377 BC), known as the father of medicine, used vinegar to irrigate open wounds and wrapped dressings around wounds to prevent further injury. His teachings remained unchallenged for centuries.
Galen (Roman gladiatorial surgeon, 130-200 AD) was first to recognize that pus from wounds inflicted by the gladiators heralded healing (pus bonum et laudabile ["good and commendable pus"]). Unfortunately, this observation was misinterpreted, and the concept of pus preempting wound healing persevered well into the eighteenth century. The link between pus formation and healing was emphasized so strongly that foreign material was introduced into wounds to promote pus formation-suppuration.
The concept of wound healing remained a mystery, as highlighted by the famous saying by Ambroise Paré (French military surgeon, 1510-1590), "I dressed the wound. God healed it" (Cohen, 1998).
Koch (Professor of Hygiene and Microbiology, Berlin, 1843-1910) first recognized the cause of infective foci as secondary to microbial growth in his nineteenth century postulates.
Semmelweis (Austrian obstetrician, 1818-1865) demonstrated a 5-fold reduction in puerperal sepsis by hand washing between performing postmortem examinations and entering the delivery room.
Joseph Lister (Professor of Surgery, London, 1827-1912) and Louis Pasteur (French bacteriologist, 1822-1895) revolutionized the entire concept of wound infection. Lister recognized that antisepsis could prevent infection (Lister, 1867). In 1867, he placed carbolic acid into open fractures to sterilize the wound and prevent sepsis and hence the need for amputation. In 1871, Lister began to use carbolic spray in the operating room to reduce contamination. As late as the nineteenth century, aseptic surgery was not routine practice. Sterilization of instruments began in the 1880s as did the wearing of gowns, masks, and gloves. Penicillin was first used clinically in 1940 by Howard Florey. With the use of antibiotics, a new era in the management of wound infections commenced.
A survey sponsored by the World Health Organization demonstrated a prevalence of nosocomial infections varying from 3-21%, with wound infections accounting for 5-34% of the total (Mayon-White, 1988).
The 2002 survey report by the National Nosocomial Infection Surveillance Service (NNIS), which covers the period between October 1997 and September 2001, indicates that the incidence of hospital acquired infection related to surgical wounds in the United Kingdom is as high as 10% and costs the National Health Service in the United Kingdom approximately 1 billion pounds (1.8 billion dollars) annually.
Collected data on the incidence of wound infections probably underestimate true incidence because most wound infections occur when the patient is discharged from hospital (about 30-40%), and these infections may be treated in the community without hospital notification.
Post-operative wound infections are the most common serious complications of surgery. It remains a major clinical problem in terms of morbidity, mortality (Astaneau et al 2001), postoperative hospital stay and hospital costs (Green et al, 1977). Based on National Nosocomial Infection Surveillance (NNIS ) system reports, SSIs are the third most frequently reported nosocomial infection, accounting for 14%-16% of all nosocomial infection among hospitalized patients (Mangram et a l, 1999). Among surgical patients, SSIs were the most common nosocomial infection, accounting for 38% of all such infections. Infection rates in US National Nosocomial Infection Surveillance system hospitals were reported to be: clean 2.1%, clean-contaminated 3.3%, contaminated 6.4% and dirty 7.1% (Culver DH et al 1991).
SSIs are associated with increased morbidity and mortality. Seventy-seven percent of the deaths of surgical patients were related to surgical wound infection (Mangram et a l, 1999).
Therefore, the purpose of the present paper is to identify the risk factors affecting the abdominal surgical site infections and their incidence employing the technique of a logistic regression model.

DATA COLLECTION AND METHODOLOGY

The study was carried out in the Department of Surgery, Enam Medical College Hospital, Sava, Dhaka, Bangladesh. Data were collected through pre- and post-operative examinations. The subjects were followed till discharge. Data were collected in standardized data collection form. All data were entered into statistical package for social science (SPSS) software for statistical analysis. Some quantitative variables have been redefined in classifications: age has been dichotomized as older than, or younger than 50 years. Weight and height were used to calculate the body mass index (BMI) (calculated as the weight in kilograms divided by the height in meters squared). According to BMI, patients were classified into three nutritional statuses: obese, normal and underweight. Obesity was defined as a BMI of 30 according to the new World Health Organization's classification. Underweight was defined as a BMI of 18.5 or less.
Types of intervention were dichotomized as abdominal versus extra-abdominal procedure. The univariate analysis was tested using Student's t-test for continuous variables and the Chi-square test for categorical variables.
Multivariate analysis was done using a logistic regression model. To test the independence of the risk factors, the significant variables (p <.05) in the univariate analyses were entered into a stepwise logistic regression equation using SPSS software to evaluate the risk of each factor when adjusted for other factors.
Patients (n = 307) who had undergone elective surgery were studied and the relationships among variables were analyzed by Student's t and Chi-square tests. To test the independence of the risk factors, the significant variables (p <.05) were entered into a stepwise logistic regression analysis.
The independent risk factors analyzed were divided into preoperative and intraoperative variables. The preoperative variables were as follows: age, sex, height, weight, the presence of diabetes mellitus, chronic liver disease, chronic renal failure, ASA score, smoking habit, hair removal, preoperative bath and preoperative hospital stay.
The intra-operative variables included the following: name of operation, length of abdominal incision, incision on a preexistent abdominal scar, perioperative blood transfusion, level of surgeon, surgical wound class, type and duration of drain, operation serial number, and the length of operative time. Dependent variables included the following: development of surgical site infection, further treatment required and postoperative hospital stay.
Bio-specimen study was started in the pre-anaesthetic checkup room. Preoperative data was collected from response to a preformed questionnaire. ASA score was collected with the help of the anaesthesiology team. ASA score was dichotomized as ASA class 1 or 2 versus ASA class 3, 4 or 5. Smoking habit was dichotomized as nonsmoker or cessation for >1 month versus smoker or cessation for <1 month.
Intra-operative data was collected by in person presence in the operating theatre during the operation. When more than one procedure was performed during the surgical intervention, the main surgical procedure was considered for analysis. Surgeons having post graduate degree in surgery were considered as a consultant.
Duration of operation is time in minutes from skin incision to skin closure. Operative time has been divided into 3 classifications: 60 minutes, from 61 to 120 minutes, and 121 minutes or longer.
Surgical site was examined on the third postoperative day and every three days thereafter till discharge of patient from hospital. The observation schedule was increased to more frequent intervals when surgical site had shown any sign of infection. The CDC NNIS definition was followed to define surgical site infection. Bacteriological culture and sensitivity test of fluid or tissue from incisional site /organ /space was performed as and when required. Infection occurring after discharge was not surveyed.

RESULTS AND DISCUSSION

In the study period 309 patients from surgery unit 1 & 2 were included for investigation. Later, two patients were withdrawn as they were discharged from hospital on the second postoperative day. Thus, 307 patients were finally studied. Among them 27 (8.8%) patients developed post-operative wound infection.

Mean age of patients is 41.9 (SD ±16.4) years, with a range of 14 - 90 years. Among them 232 patients (75.5%) are of 50 years or younger. 75 (24.5%) patients' ages are more than 50 years. Wound infection rate is significantly higher in the older age group with a p value of <0.003. In the study 56.7% (174) of the patients were male. Post operative wound infection rate is a bit higher in males (9.8% versus 7.5%), but the difference is not statistically significant.
In the study, fifteen patients (4.9%) were diabetic. Among them 4 (26.7%) patients developed post operative wound infection. This observation is statistically highly significant with a p value of <0.02. In the study, 295 (96.1%) patients had ASA score of 1 or 2 and 12 (3.9%) patients had ASA score of >3. Among the 12 patients with ASA score of >3, 5 (41.7%) patients developed post operative wound infection. This observation indicates that the post operative wound infection rate is significantly higher in patients with ASA score of ?3 with a p value of <0.0001. Hair on the skin over the operative site was either not removed (47.2%) or was removed the day before surgery (30.9%) in the majority of patients. Only 67 (21.8%) patients had the preoperative area shaved just before surgery, contrary to the current CDC guidelines. Post operative wound infection rate was 6% in patients who shaved prior to surgery, 9% in patients who did not shave and 10.5% among those who shaved day before operation. The difference in SSI rate was statistically insignificant with p value of >0.5. In this study, most of the operations (87.3%), were done by consultant surgeon. SSI rate was unusually higher when done by consultants, but the difference is not significant (p>0.7). In 53 (17.3%) patients, incisions were given on a pre-existing scar. SSI rate was lower when incision was given on a pre-existing scar. But the difference is insignificant with a p value of >0.3.
The majority of operations (170) took less than 60 minutes from skin incision to skin closure: 23 (26.1%) operation required more than 2 hours to be completed. The duration of surgical operation also proved to be a significant factor: only 5.3% of operations lasting 60 minutes or less led to infection, while for operations lasting more than 2 hours this rate leapt to 26.1%.

All the risk factors for SSI with p value < 0.05 in univariate analysis were entered into a stepwise logistic regression model for multivariate analysis. ASA score (p<.0001), Diabetes mellitus (p<.004), duration of operation (p<.004), and older age group (p<.006) were proved to be an independent risk factor for wound infection.

a Predictors: (Constant), ASA score
b Predictors: (Constant), ASA score, Diabetes mellitus
c Predictors: (Constant), ASA score, Diabetes mellitus, Operation time
d Predictors: (Constant), ASA score, Diabetes mellitus, Operation time, Age group
e Dependent Variable: Wound Infection

CONCLUSSION AND RECOMMENDATION

This study provided information on rate and risk factors for SSI occurrence in elective operations in the Department of General Surgery of Enam Medical College Hospital, Savar, Dhaka.
To get information on SSI rate, further study on a large scale is needed, including all patients in the study population. Four independent risk factors were identified in both univariate and multivariate analysis. Another five risk factors were identified in univariate analysis but not in multivariate analysis. With the aim of reducing the rate of infectious complications, the risk factors can be divided into the following two categories:

• Unmodifiable factors: age, ASA score, class 3 surgical site and peri operative blood transfusion.
  
• Factors that are able to be modified before or during surgery: being underweight, preoperative hospital stay, duration of drain for more than three days, diabetes mellitus and operative time.
  

Could correction of the modifiable factors reduce the postoperative infectious complications? Further study is required to obtain the answer. In conclusion, SSI surveillance should be conducted and maintained in all hospitals to promote better surgical outcomes. The following recommendations are made for consideration:

• We should do our best to reduce the average operation duration to less than 2 hours.
  
• The average preoperative bed stay should be reduced.
  
• The time of shaving should approximate the operation time as much as possible.
  
• Drain should be withdrawn as soon as it is no longer aspirate.
  
• Good control of glycaemic status should be achieved.
  
• Enteral feeding should be resumed as soon as possible in pre and postoperative period.
  
• The efficacy of these proposals should be evaluated by a prospective study.
  

REFERENCES

Astaneau P, Rioux C, Brucker G. Morbidity and mortality associated with surgical site infections: results from the 1997-1999 INCISO surveillance. J Hosp Infect. 2001; 48: 267-74.

Breasted D., 1930.The Edwin Smith Surgical Papyrus. University of Chicago. University of Chicago press; USA.

Bryan PW.,1930. The Papyrus Ebers. London/Washington DC: Government Printing Office .

Cohen IK. 1998. A brief history of wound healing. 1st ed. Yardley, Pa: Oxford Clinical Communications Inc. 1998.

Culver DH, Horan TC, Gaynes RP, et al. Surgical wound infection rates by wound class, operative procedure, and patient risk index: National Nosocomial Infections Surveillance System. Am J Med. 1991; 91 (suppl 3B): 152S-157S.

Green J W and R P Wenzel. Postoperative wound infection: a controlled study of the increased duration of hospital stay and direct cost of hospitalization. Ann Surg. 1977; 185(3): 264-268.

Lister J. 1867. On a new method of treating compound fractures. Lancet. Vol.1, pp : 326-329, 387-389, 507-509.

Mayon-White RT., Ducel G., Kereselidze T.,1988 . An international survey of the prevalence of hospital-acquired infection. J Hosp Infect. 1988; 11: 43-4.

Mangram AJ, Horan CT, Pearson ML. Guideline for prevention of surgical site infection, 1999. Inf Cont Hosp Epid, 1999; 20(4): 247-278.