Influence of Operating Room Surface Contamination on Surgical Wounds A
Prospective Study
Donald O. Weber, MD; James J. Gooch, MPH; Walter R. Wood, MA Stat; Eugene M. Britt, PhD; Richard O. \s=b\ The influence of operating room contamination on wound infection rates in clean, clean-contaminated, contaminated, and septic procedures was studied by a prospective randomized study of 2,020 surgical wounds. Operating room surface contamination was assessed by the RODAC bacterial plate method. Control rooms uniformly received Wet-Vac cleaning between operations. Experimental rooms were not cleaned between consecutive clean operations, but were cleaned after contaminated operations. The difference in surface contamination between groups of experimental and control rooms was found to be significant at the P < .05 level. Patients operated on in experimental and control rooms were followed up postoperatively to assess whether they experienced wound infection. No statistically significant differences in wound infection rates were found between experimental and control room operations as total groups, clean procedures, or operations of long duration.
(Arch Surg 111:484-488, 1976) infections are an all too frequent complication of surgery, and result in significant mortality and morbidity. Economic loss from wound infections in the United States has been estimated by Alexander' to exceed
Wound
Kraft,
MD
$9.4 billion annually. Critical attention has focused
on the since studies indicate that wounds destined to become infected experience the responsible contamination during the operation.7 The source of these bacteria is either endogenous (patient related) or exoge¬ nous (environmental), with the former the more frequent factor.4 The most conspicuous source of exogenous contamina¬ tion has been the operating room itself. Improved surface cleaning techniques have been proposed to reduce this factor.7"' Investigations by others imply that sources of contamination ultimately contributing to wound infection are removed by vigorous theater cleaning.7'1" However, the definitions of adequate cleaning are based on laboratory standards only." A search of the surgical and epidemio¬ logica! literature fails to discover a study correlating directly measured operating room surface contamination with the incidence of postoperative wound infection. A prospective study was initiated to answer this critical question: what influence, if any, does the universally recommended between-each-operation floor cleaning have on the frequency of infection in surgical wounds?
operating
room,
'
,7>
MATERIALS AND METHODS
Accepted
publication Dec 4, 1975. departments of general surgery (Drs Weber and Kraft) and pathology (Drs Gooch and Britt), St Joseph Mercy Hospital, Ann Arbor, and Eastern Michigan University, Ypsilanti, Mich (Mr Wood). Read before the 83rd annual meeting of the Western Surgical Association, Colorado Springs, Colo, Nov 22, 1975. Reprint requests to Department of General Surgery, St Joseph Mercy Hospital, 326 N Ingalls St, Ann Arbor, MI 48104 (Dr Kraft). From the
for
Seven arbitrarily selected operating rooms at St Joseph Mercy Hospital composed the study group, and received their usual quota of general and subspecialty surgical cases each day. Rooms were selected randomly, with new control and experimental groups chosen monthly to avoid unsuspected intrinsic factors. Treatment of horizontal surfaces (operating room table, Mayo stand, and side shelves) was the same for both experimental and
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/10/2015
control groups. Before the first operation each day, all horizontal surfaces were wet with a germicidal detergent, Ves-Phene, allowed to stand five minutes, then wiped and allowed to air dry. This sequence was repeated in all rooms between each opera¬ tion. The control and experimental groups differed in the choice of treatment given the operating room floor. In the control group, before the first operation the entire floor surface was wet with Ves-Phene (one ounce per one gallon of water), allowed to stand five minutes, and then finished with Wet-Vac pickup. This same treatment was performed between each operation, and after the last operation of the day. In the experimental group, floors were untouched before the first operation, having been cleaned the night before. Treatment between operations was a graded one based on the degree of contamination related to the previous procedure, as subdivided by the National Research Council (1964)'3: clean, no known contami¬ nation, and gastrointestinal, genitourinary, and respiratory tracts
entered; clean-contaminated, gastrointestinal, genitourinary, respiratory tracts entered but no spillage; contaminated, trauma: gross spillage from hollow viscus or acute inflammation; and septic, purulence or uncontrolled perforated viscus present. Floors of experimental group rooms following operations clas¬ sified as contaminated or septic were given full-floor treatment identical to that in control rooms. Experimental group floors following operations classified as clean or clean-contaminated received no treatment unless gross soilage occurred (eg, bloody sponge), and then only local cleaning was done with a towel and not
and
Ves-Phene. Only those
operations between 7:00 am and the last scheduled operations, Monday through Friday, were included in the study. Complete Wet-Vac treatment was given relative to emergency operations during evening hours and weekends. Assignment of emergency operations to the rooms was random.
Bactériologie Monitoring widespread method to evaluate adequacy of cleaning is the plates.17 Colony counts between 10 and 25/25 sq cm are currently accepted by the American Public Health Association A
use
of RODAC
(APHA)" and the Center for Disease Control (CDC).'7 These criteria are based on sample surfaces given vigorous cleaning with Wet-Vac technique before traffic is permitted. One room from the control group and one room from the experimental group were selected randomly and monitored weekly on a random day. Both rooms were monitored before the first operation of the day, and a second time in the afternoon at an unexpected hour. The morning and afternoon counts were average for each of the two rooms monitored. Thirty RODAC plates were placed in each room at each sample time, 15 within randomly selected areas on the operating room floor, and 15 on horizontal surfaces. The RODAC plates were counted at 48 hours' incubation at 35 C.'7 Colonies greater than 200/25 sq cm were considered confluent growth. All monitoring and counting was performed by one person. All floor and horizontal surface treatments were performed by one person during 41 of the 42 weeks of the study. The cleaner was "blind" to the day and room of RODAC monitor¬
ing.
HYPOTHESES
The following hypotheses were proposed at the inception of the investigation so that the study would be prospec¬ tive. 1. The different cleaning methods will result in different levels of contamination on surfaces, as measured by RODAC plates placed on floors and horizontal surfaces. 2. Postoperative wound infection rates will not be increased statistically by the alteration in operating room
cleaning technique.
3. Postoperative wound infection rates will not be increased in groups of patients undergoing clean oper¬ ations in rooms prepared by the altered technique. 4. Postoperative wound infection rates will not be increased in groups of patients experiencing operations longer than three hours in experimentally cleaned rooms.
RESULTS
Daily contamination for control and experimental rooms estimated by averaging morning and afternoon bacte¬
was
rial counts. Floor and horizontal surface bacterial counts were analyzed separately. Experimental and control counts were paired by the day of sampling, and statistical tests were performed on these paired differences." For the floors, the mean bacterial count in experimental rooms was 11.70 colonies/25 sq cm, in the control rooms, 2.24 colonies/25 sq cm. The mean difference was 9.46. The standard error of the difference was .8783, yielding an observed t value of 10.77. The difference in bacterial counts between control and experimental operating room floors was significant at the < .05 level. For the horizontal surfaces, the mean bacterial count in experimental rooms was 6.98 colonies/sq cm, and in control rooms, 5.86 colonies/sq cm. The mean difference was 1.12, with a standard error of the difference of .7778, yielding an observed t value of 1.45. The difference in bacterial counts between control and experimental horizontal surfaces was found to be significant at the < .08 level.
Postoperative Wound Infections A total of 2,020
patients were followed up; complete data four dichotomous variables were collected on 2,013. These variables were (1) treatment: the procedure was performed either in a control or experimental operating room; (2) case classification: the procedure was classed as clean, clean-contaminated, contaminated, or septic, with regard to endogenous contamination factors; (3) length: the procedure was either short (less than three hours) or
on
Wound Infection Data Patients operated on were followed up 14 days postoperatively assess the presence or absence of wound infection. The defini¬ tion of the National Research Council was used: "Any surgical wound which drains purulent material, with or without a positive culture, is considered the source of a nosocomial infection. The to
of the organism, whether endogenous or exogenous, is not considered."17 Procedures not resulting in an observable wound (dilation and curettage, bronchoscopy, etc) were excluded from the study. Wound infections detected in the hospital were reported to the Infection Control Committee. All patients discharged were contacted via questionnaire designed to elicit false-positive rather than false-negative replies. Patients not responding, as well as those reporting infections, were contacted by telephone to verify whether or not the definition criterion was met. Data were stored on magnetic tape using a digital computer for the 42 weeks of the study, and examined only at termination of the project. source
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/10/2015
Table
1.—Follow-up of Postoperative Wound No. of Patients (N Control 1,057
Treatment
Case classification*
Not clean 279
Lengtht Proportion
infected
=
Infections
2,013) Experimental 956
Clean 778
Short
Long
227 0.0837
52 0.2115
Short 712 0.0295
Clean 756
Not clean 200
Short 160 0.1125
Long 66 0.1061
Long
Short
Long
40 0.1000
719
37
0.0223
Clean indicates clean operations only; not clean indicates clean-contaminated, contaminated and septic operations, t Short indicates operations of less than three hours; long indicates operations of three or more hours. *
Table 2.—Statistical Variable
Analysis
60
Standard Error 0.20579
Critical Ratio -6.870
NSf
Mean
Effects* -1.4137
Length of procedure
+0.0058
0.19960
+0.029
-0.3479
0.07622
-4.564
Prospective Study
Donald O. Weber, MD; James J. Gooch, MPH; Walter R. Wood, MA Stat; Eugene M. Britt, PhD; Richard O. \s=b\ The influence of operating room contamination on wound infection rates in clean, clean-contaminated, contaminated, and septic procedures was studied by a prospective randomized study of 2,020 surgical wounds. Operating room surface contamination was assessed by the RODAC bacterial plate method. Control rooms uniformly received Wet-Vac cleaning between operations. Experimental rooms were not cleaned between consecutive clean operations, but were cleaned after contaminated operations. The difference in surface contamination between groups of experimental and control rooms was found to be significant at the P < .05 level. Patients operated on in experimental and control rooms were followed up postoperatively to assess whether they experienced wound infection. No statistically significant differences in wound infection rates were found between experimental and control room operations as total groups, clean procedures, or operations of long duration.
(Arch Surg 111:484-488, 1976) infections are an all too frequent complication of surgery, and result in significant mortality and morbidity. Economic loss from wound infections in the United States has been estimated by Alexander' to exceed
Wound
Kraft,
MD
$9.4 billion annually. Critical attention has focused
on the since studies indicate that wounds destined to become infected experience the responsible contamination during the operation.7 The source of these bacteria is either endogenous (patient related) or exoge¬ nous (environmental), with the former the more frequent factor.4 The most conspicuous source of exogenous contamina¬ tion has been the operating room itself. Improved surface cleaning techniques have been proposed to reduce this factor.7"' Investigations by others imply that sources of contamination ultimately contributing to wound infection are removed by vigorous theater cleaning.7'1" However, the definitions of adequate cleaning are based on laboratory standards only." A search of the surgical and epidemio¬ logica! literature fails to discover a study correlating directly measured operating room surface contamination with the incidence of postoperative wound infection. A prospective study was initiated to answer this critical question: what influence, if any, does the universally recommended between-each-operation floor cleaning have on the frequency of infection in surgical wounds?
operating
room,
'
,7>
MATERIALS AND METHODS
Accepted
publication Dec 4, 1975. departments of general surgery (Drs Weber and Kraft) and pathology (Drs Gooch and Britt), St Joseph Mercy Hospital, Ann Arbor, and Eastern Michigan University, Ypsilanti, Mich (Mr Wood). Read before the 83rd annual meeting of the Western Surgical Association, Colorado Springs, Colo, Nov 22, 1975. Reprint requests to Department of General Surgery, St Joseph Mercy Hospital, 326 N Ingalls St, Ann Arbor, MI 48104 (Dr Kraft). From the
for
Seven arbitrarily selected operating rooms at St Joseph Mercy Hospital composed the study group, and received their usual quota of general and subspecialty surgical cases each day. Rooms were selected randomly, with new control and experimental groups chosen monthly to avoid unsuspected intrinsic factors. Treatment of horizontal surfaces (operating room table, Mayo stand, and side shelves) was the same for both experimental and
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/10/2015
control groups. Before the first operation each day, all horizontal surfaces were wet with a germicidal detergent, Ves-Phene, allowed to stand five minutes, then wiped and allowed to air dry. This sequence was repeated in all rooms between each opera¬ tion. The control and experimental groups differed in the choice of treatment given the operating room floor. In the control group, before the first operation the entire floor surface was wet with Ves-Phene (one ounce per one gallon of water), allowed to stand five minutes, and then finished with Wet-Vac pickup. This same treatment was performed between each operation, and after the last operation of the day. In the experimental group, floors were untouched before the first operation, having been cleaned the night before. Treatment between operations was a graded one based on the degree of contamination related to the previous procedure, as subdivided by the National Research Council (1964)'3: clean, no known contami¬ nation, and gastrointestinal, genitourinary, and respiratory tracts
entered; clean-contaminated, gastrointestinal, genitourinary, respiratory tracts entered but no spillage; contaminated, trauma: gross spillage from hollow viscus or acute inflammation; and septic, purulence or uncontrolled perforated viscus present. Floors of experimental group rooms following operations clas¬ sified as contaminated or septic were given full-floor treatment identical to that in control rooms. Experimental group floors following operations classified as clean or clean-contaminated received no treatment unless gross soilage occurred (eg, bloody sponge), and then only local cleaning was done with a towel and not
and
Ves-Phene. Only those
operations between 7:00 am and the last scheduled operations, Monday through Friday, were included in the study. Complete Wet-Vac treatment was given relative to emergency operations during evening hours and weekends. Assignment of emergency operations to the rooms was random.
Bactériologie Monitoring widespread method to evaluate adequacy of cleaning is the plates.17 Colony counts between 10 and 25/25 sq cm are currently accepted by the American Public Health Association A
use
of RODAC
(APHA)" and the Center for Disease Control (CDC).'7 These criteria are based on sample surfaces given vigorous cleaning with Wet-Vac technique before traffic is permitted. One room from the control group and one room from the experimental group were selected randomly and monitored weekly on a random day. Both rooms were monitored before the first operation of the day, and a second time in the afternoon at an unexpected hour. The morning and afternoon counts were average for each of the two rooms monitored. Thirty RODAC plates were placed in each room at each sample time, 15 within randomly selected areas on the operating room floor, and 15 on horizontal surfaces. The RODAC plates were counted at 48 hours' incubation at 35 C.'7 Colonies greater than 200/25 sq cm were considered confluent growth. All monitoring and counting was performed by one person. All floor and horizontal surface treatments were performed by one person during 41 of the 42 weeks of the study. The cleaner was "blind" to the day and room of RODAC monitor¬
ing.
HYPOTHESES
The following hypotheses were proposed at the inception of the investigation so that the study would be prospec¬ tive. 1. The different cleaning methods will result in different levels of contamination on surfaces, as measured by RODAC plates placed on floors and horizontal surfaces. 2. Postoperative wound infection rates will not be increased statistically by the alteration in operating room
cleaning technique.
3. Postoperative wound infection rates will not be increased in groups of patients undergoing clean oper¬ ations in rooms prepared by the altered technique. 4. Postoperative wound infection rates will not be increased in groups of patients experiencing operations longer than three hours in experimentally cleaned rooms.
RESULTS
Daily contamination for control and experimental rooms estimated by averaging morning and afternoon bacte¬
was
rial counts. Floor and horizontal surface bacterial counts were analyzed separately. Experimental and control counts were paired by the day of sampling, and statistical tests were performed on these paired differences." For the floors, the mean bacterial count in experimental rooms was 11.70 colonies/25 sq cm, in the control rooms, 2.24 colonies/25 sq cm. The mean difference was 9.46. The standard error of the difference was .8783, yielding an observed t value of 10.77. The difference in bacterial counts between control and experimental operating room floors was significant at the < .05 level. For the horizontal surfaces, the mean bacterial count in experimental rooms was 6.98 colonies/sq cm, and in control rooms, 5.86 colonies/sq cm. The mean difference was 1.12, with a standard error of the difference of .7778, yielding an observed t value of 1.45. The difference in bacterial counts between control and experimental horizontal surfaces was found to be significant at the < .08 level.
Postoperative Wound Infections A total of 2,020
patients were followed up; complete data four dichotomous variables were collected on 2,013. These variables were (1) treatment: the procedure was performed either in a control or experimental operating room; (2) case classification: the procedure was classed as clean, clean-contaminated, contaminated, or septic, with regard to endogenous contamination factors; (3) length: the procedure was either short (less than three hours) or
on
Wound Infection Data Patients operated on were followed up 14 days postoperatively assess the presence or absence of wound infection. The defini¬ tion of the National Research Council was used: "Any surgical wound which drains purulent material, with or without a positive culture, is considered the source of a nosocomial infection. The to
of the organism, whether endogenous or exogenous, is not considered."17 Procedures not resulting in an observable wound (dilation and curettage, bronchoscopy, etc) were excluded from the study. Wound infections detected in the hospital were reported to the Infection Control Committee. All patients discharged were contacted via questionnaire designed to elicit false-positive rather than false-negative replies. Patients not responding, as well as those reporting infections, were contacted by telephone to verify whether or not the definition criterion was met. Data were stored on magnetic tape using a digital computer for the 42 weeks of the study, and examined only at termination of the project. source
Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/10/2015
Table
1.—Follow-up of Postoperative Wound No. of Patients (N Control 1,057
Treatment
Case classification*
Not clean 279
Lengtht Proportion
infected
=
Infections
2,013) Experimental 956
Clean 778
Short
Long
227 0.0837
52 0.2115
Short 712 0.0295
Clean 756
Not clean 200
Short 160 0.1125
Long 66 0.1061
Long
Short
Long
40 0.1000
719
37
0.0223
Clean indicates clean operations only; not clean indicates clean-contaminated, contaminated and septic operations, t Short indicates operations of less than three hours; long indicates operations of three or more hours. *
Table 2.—Statistical Variable
Analysis
60
Standard Error 0.20579
Critical Ratio -6.870
NSf
Mean
Effects* -1.4137
Length of procedure
+0.0058
0.19960
+0.029
-0.3479
0.07622
-4.564
