Post-operative infection of endoscopic submucosal dissection of early colorectal neoplasms: a case–controlled study using a Japanese database
T. Muro* PhD, N. Higuchi* PhD, M. Imamura* PhD, H. Nakagawa* PhD, M. Honda† PhD, K. Nakao‡ MD PhD, K. Izumikawa§ MD PhD, H. Sasaki* PhD and T. Kitahara* PhD *Department of Clinical Pharmacy, Nagasaki University Hospital, †Department of Medical Information, Nagasaki University Hospital, ‡Department of Gastroenterology and Hepatology, Nagasaki University Hospital, and §Infection Control and Education Center, Nagasaki University Hospital, Nagasaki, Japan
Received 19 February 2015, Accepted 2 July 2015
Keywords: antibiotic prophylaxis, pharmacoepidemiology, surgical wound infection
What is known and objective: Endoscopic submucosal dissection of early colorectal neoplasms (ESD-ECN) is known to be an operation with risk of contamination, possibly requiring pre-operative antimicrobial prophylaxis for the prevention of post-operative infection. However, an evaluation of the need for pre-operative antimicrobial prophylaxis for ESD-ECN has yet to be reported. The objective of this study was to determine whether pre-operative antimicrobial prophylaxis is associated with a reduced incidence of post-operative infection following ESD-ECN. Methods: The present retrospective case–controlled study uti- lized a database built from the medical records of 14 university hospitals throughout Japan. Patients who were admitted and discharged from the hospital from April 2012 to October 2013 and who had undergone ESD-ECN were included in the study. Patients who had been undergone any other operation during their course of hospitalization, and patients who were prescribed antimicrobial agents for reasons other than post-operative infection or for prophylaxis were excluded. Characteristics of the study population, pre-operative antimicrobial prophylaxis and antimicrobial therapy for post-operative infection were investigated. In addition, we compared the characteristics of patients with post-operative infection (PI) and those with no post-operative infection (NPI). Univariate analyses were used to estimate the odds ratios (OR) and 95% confidence intervals (95% CI). Results and discussion: We obtained the records of 522 patients who had undergone ESD-ECN from the database. After appli- cation of exclusion criteria, 421 patients were enrolled. The post- operative infection rate was 1�2%. Peritonitis was found most to be the most common post-operative infection (44%). Pre-opera- tive antimicrobial prophylaxis was used for 314 patients (75%), with a median duration of 3�0 (range 1–11) days. Cefotiam was most frequently prescribed for pre-operative antimicrobial prophylaxis (56%). Antimicrobial therapies were started 1–10 days after ESD-ECN for a duration of 1–14 days. Pre- operative antimicrobial prophylaxis was not associated with post-operative infection rate, with an OR (95% CI) of 0�73 (0�08–6�61). However, digestive tract perforation was shown to be
associated with post-operative infection and had an OR (95% CI) of 17�1 (1�66–176�45). What is new and conclusion: Post-operative infection is an exceedingly rare event following ESD-ECN.Pre-operative antimi- crobial prophylaxis had no significant effect on post-operative infection following ESD-ECN and thus may be unnecessary. Instead, prevention of digestive tract perforation may be more critical for the decrease in post-operative infections.
WHAT IS KNOWN AND OBJECTIVE
Endoscopic submucosal dissection (ESD) is a variant of endoscopic mucosal resection (EMR) and is a technique for resection of early gastrointestinal tract cancers. ESD is used primarily in the stomach, but is also increasingly being applied in the colon and rectum.1,2 The technique involves identifying the margins of the polyp, submucosal injection and circumferential dissection of the tumour-bearing mucosa and submucosa using various diathermic knives. The aim of ESD is to achieve an en bloc resection of a sessile lesion, irrespective of its size. This facilitates histological evaluation and improves the rate of curative resection, even for carcinomas with early submucosal invasion.3
Post-operative infections are a major contributor to healthcare- associated infections. Gastrointestinal procedures are thought to have the highest risk for post-operative infection due to exposure to intraluminal bacteria.4 Therefore, pre-operative antimicrobial prophylaxis is warranted for patients undergoing gastrointestinal surgery involving the colon or rectum. ESD of early colorectal neoplasms (ESD-ECN) is considered to be an operation with risk of contamination, possibly requiring pre-operative antimicrobial prophylaxis for the prevention of post-operative infection. Because the ESD procedure requires advanced skill and extensive training to achieve a satisfactory level of ability, the use of this technique has been somewhat limited. Therefore, the evaluation of the need for pre-operative antimicrobial prophylaxis for ESD has not been sufficient. The objective of this study was to determine whether pre-operative antimicrobial prophylaxis is associated with a reduced incidence of post-operative infection following ESD-ECN.
Study design and data source
This study was a retrospective case–controlled study utilizing the Platform for Clinical Information Statistical Analysis (CISA)
Correspondence: T. Muro, Department of Clinical Pharmacy, Nagasaki University Hospital, 1-7-1 Sakamoto, Nagasaki, Nagasaki 852-8501, Japan. Tel.: +81 95 819 7249; fax: +81 95 819 7251; e-mail: muroth1@ nagasaki-u.ac.jp
© 2015 John Wiley & Sons Ltd 573
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 573–577 doi: 10.1111/jcpt.12313
database, which is composed of clinical data from 14 university hospitals throughout Japan. The CISA database contains data on approximately 2�45 million unique patients, a cumulative total of 1�25 million inpatients, a cumulative total of 37�79 million outpatients, 32�86 million prescriptions for inpatients, and 14�50 million prescriptions for outpatients. These data were collected from the medical records from each facility after removing personal information and irreversibly anonymizing the data. At present, data on medical treatment results, such as test results and interviews, information on palpation, and images are not included in the database. Diagnostic codes were established according to both the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10) and the Japanese Receipt Disease Name Master (JRDNM). Drug codes were estab- lished according to the Anatomical Therapeutic Chemical Classi- fication System (ATC) code. In this study, only data obtained from April 2012 to October 2013 were analysed.
Population and definitions
Data on patients who were admitted and discharged from the hospital from April 2012 to October 2013 and who had undergone ESD-ECN were collected from the CISA database. To reduce any possible influence of the patients’ pre-operative conditions, patients who had undergone ESD-ECN more than 4 days after the date of admission were excluded. A duration of 4 days was selected because the duration was the most generally a period from hospitalization to do ESD-ECN in Japan. Patients who had undergone any other operation other than their first ESD-ECN procedure during their course of hospitalization, such as those who had undergone additional tumorectomy or ESD for the same or different region, were also excluded. Patients who were prescribed antimicrobial agents for treatment of conditions other than post-operative infection were excluded, as were patients who
were prescribed antimicrobial agents for purposes other than post- operative infection or prophylaxis. The patients remaining after application of the exclusion criteria were divided into two groups, those who had been post-operative infection (PI) and those who had been not PI (NPI) (Fig. 1).
Codes from the Japanese Receipt Disease Name Master were used for the definition of post-operative infection. Codes for infectious diseases were selected by excluding the infections originating from regions other than the abdomen, such as pneumonia, from all disease codes of the study population, and defined the remaining as Post-operative Infection Codes (PIC). In this study, antimicrobial therapy for post-operative infection was defined as antimicrobial therapy administered to patients who were assigned a PIC after undergoing ESD-ECN. Thus, the post- operative infection (PI) group was defined as patients who were treated with antimicrobial therapy for post-operative infection.
Pre-operative antimicrobial prophylaxis was defined as antibi- otic administration during the first day of hospitalization to the day of ESD-ECN.
Epidemiological research and statistical methods
Characteristics of the study population, use of pre-operative antimicrobial prophylaxis and administration of antimicrobial therapy for post-operative infection were collected from the database for epidemiological evaluation. JRDNM code was used for definition of the perforation of the digestive tract. Perforation of the digestive tract was defined as patients who were assigned JRDNM codes that were meaning ‘perforation of the digestive tract’ in the disease name after undergoing ESD-ECN and counted. Additionally, the duration of hospitalization after ESD-ECN and the number of patients with malignant tumour-related diseases were collected for comparison between the PI and NPI groups. We compared the PI and NPI groups using the Wilcoxon rank sum test
Endoscopic submucosal dissection of early colorectal neoplasms (ESD – ECN) (n = 522)
Post-operative infection group (PI) (n = 5)
No – post-operative infection group (NPI) (n = 416)
ESD – ECN was performed more than 4 days after date of admission (n = 61)
Patients who had undergone operations other than ESD – ECN (including additional tumorectomy or ESD for the same or different region) (n = 13)
Patients who were prescribed antimicrobial agents for purposes other than post-operative infection or prophylaxis (n = 27)
Fig. 1. Flow chart for selection of study population.
© 2015 John Wiley & Sons Ltd Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 573–577 574
Post-operative infection of endoscopic T. Muro et al.
for continuous variables, the chi-square, or Fisher’s exact test for dichotomous variables. Differences were considered significant when P < 0�05. Univariate analyses was used to determine independent predictors of post-operative infection and to obtain odds ratios (OR), and the 95% confidence interval (95% CI) for each OR was calculated. Statistical significance was determined by 95% confidence intervals, not including 1�00 for logistic analyses. We introduced pre-operative antimicrobial prophylaxis and diges- tive tract perforations as groups of variables into the model, because these are well known to be associated with post-operative infection.
Because the data utilized in this study were already anonymized by the database provider, CISA, the study was exempted from obtaining informed consent from individual patients according to the local ethical guidelines for epidemiological research. This study and the waiver of informed consent were approved by the Nagasaki University Hospital Ethics Committee (14102796).
RESULTS AND DISCUSSION
Characteristics of the study population
We obtained data on 522 patients who had undergone ESD-ECN from the CISA database. After applying the exclusion criteria, 421 patients were enrolled in the present study (Fig. 1). The post- operative infection rate was 1�2%. Types of neoplasms observed in the study population are shown in Table 1. The majority of the patients had malignant neoplasms of the colon (45%) or rectum (14%). Perforation of the digestive tract was observed in 2% of patients. The characteristics of post-operative infection of the study population are shown in Table 2. Peritonitis (including general- ized, circumscribed and perforative peritonitis) was the most common post-operative infection (44%). Only two cases were reported to have developed sepsis (0�5%).
Characteristics of pre-operative antimicrobial prophylaxis
Pre-operative antimicrobial prophylaxis was used for 314 patients (75%). The median dosing period of pre-operative antimicrobial prophylaxis was 3�0 (range 1–11) days. Table 3 shows the duration of the pre-operative antimicrobial prophylaxis per generic name. Cefotiam was the most commonly prescribed pre-operative antimicrobial prophylaxis (56%), with a median dosing period of 3�0 days. Cefmetazole was the second most frequently prescribed prophylaxis (17%), with a median dosing period of 2�5 days.
Antimicrobial therapy for post-operative infection
The median duration of antimicrobial therapy for post-operative infection was 4 days; the characteristics are shown in Table 4. Six agents, including combination therapy, were used for the treat- ment of post-operative infection and were started 1–10 days after ESD-ECN for a duration of 1–14 days.
Comparison of the PI and NPI groups
Characteristics of the PI and NPI groups are shown in Table 5, with no significant differences between the two groups. Duration of stay after ESD of PI was significantly longer than that of NPI
(P = 0�03). Table 6 shows the duration of the pre-operative antimicrobial prophylaxis per generic name of PI. Cefotiam was the most commonly prescribed pre-operative antimicrobial pro- phylaxis (50%) of PI, too. Univariate analyses of PI and NPI are shown in Table 7. Pre-operative antimicrobial prophylaxis was not associated with post-operative infection and had an OR (95% CI) of 0�73 (0�08–6�61). However, perforation of the digestive tract was
Table 1. Patient neoplasms
ICD-10 code Description n (%)
C18 Malignant neoplasm of colon 347 (45) C20 Malignant neoplasm of rectum 109 (14) C78 Secondary malignant neoplasm
of respiratory and digestive organs 52 (7)
C16 Malignant neoplasm of stomach 47 (6) C34 Malignant neoplasm of bronchus and lung 32 (4) C25 Malignant neoplasm of pancreas 24 (3) C79 Secondary malignant neoplasm
of other and unspecified sites 19 (2)
C22 Malignant neoplasm of liver and intrahepatic bile ducts
C61 Malignant neoplasm of prostate 14 (2) C67 Malignant neoplasm of bladder 11 (1) C80 Malignant neoplasm, without
specification of site 10 (1)
C15 Malignant neoplasm of oesophagus 9 (1) C50 Malignant neoplasm of breast 8 (1) C91 Lymphoid leukaemia 8 (1) C77 Secondary and unspecified
malignant neoplasm of lymph nodes 7 (1)
C85 Other and unspecified types of non-Hodgkin lymphoma
C19 Malignant neoplasm of rectosigmoid junction
Others 41 (5)
ICD-10, International Statistical Classification of Diseases and Related Health Problems 10th revision. Including multiple primaries and metastases.
Table 2. Post-operative infections
PIC numbera Post-operative infectionb n
389004 Sepsis 2 8833267 Diarrhoeal disease 2 91023 Enteritis 1 5679005 Generalized peritonitis 1 5679007 Circumscribed peritonitis 1 5679012 Perforative peritonitis 1 5679015 Peritonitis 1
aJapanese Receipt Disease Name Master was used for the definition of post- operative infection. bCodes for infectious diseases were selected by excluding the infections originating from regions other than the abdomen, such as pneumonia, from all disease codes of the study population, and defined the remaining as Post-operative Infection Codes (PIC). Includes complex infections.
© 2015 John Wiley & Sons Ltd Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 573–577 575
Post-operative infection of endoscopic T. Muro et al.
associated with post-operative infection and had an OR (95% CI) of 17�1 (1�66–176�45).
The present results suggested that there is no significant correlation between pre-operative antimicrobial prophylaxis and incidence of post-operative infection after ESD-ECN. Instead, perforation of the digestive tract was found to increase post- operative infection approximately 17-fold. In present study, it is not possible to perform multivariate logistic regression analysis, because the number of patients with post-operative infection is few. So we conducted univariate analyses. Therefore, our results may include the confounding factors.
The post-operative infection rate in the present study was 1�2% and was lower than the rate previously reported for whole-colon surgery (15%).5 Sepsis had developed in only 2 patients (0�5%). Similarly, Min et al. reported a low bacteremia rate (2�5%) associated with ESD or endoscopic mucosal resection (EMR),6
suggesting that this was due to the low possibility of direct injection into a blood vessel during submucosal injection.6 This may also be the case in ESD, which is usually used for resecting a large lesion, and consequently results in a considerable amount of exposed submucosa, requiring a large number of submucosal injections. Therefore, ESD may have a higher risk for post- operative infection than EMR. However, ESD has considerably fewer opportunities of injury to blood vessels compared with
Table 3. Pre-operative antimicrobial prophylaxis
ATC level5 name na (%) Duration (day), median (range)
Cefotiam 182 (56) 3�0 (1–7) Cefmetazole 56 (17) 2�5 (1–11) Ampicillin and enzyme inhibitor 46 (14) 3�0 (1–7) Meropenem 10 (3) 5�0 (4–6) Cefazolin 10 (3) 3�0 (1–3) Flomoxef 8 (2) 2�0 (2–4) Ceftriaxone 7 (2) 3�0 (1–9) Cefoperazone, combination 5 (2) 3�0 (2–9) Clindamycin 1 (0) 4�0 (–)
aIncluding combination therapy. ATC, Anatomical Therapeutic Chemical Classification System.
Table 5. Characteristics of the PI and NPI groups
PI NPI P value
N 5 416 – Age (year), median (range) 73 (54–87) 68 (54–92) 0�72 Duration of stay after ESD (day), median (range)
8 (5–22) 6 (3–20) 0�03
Male n (%) 3 (60) 245 (59) 1�00 Classification of Diseasesa
Infectious and parasitic diseases 5 119 0�91 Neoplasms 5 414 Diseases of the blood and blood-forming organs and disorders involving the immune system
Endocrine, nutritional and metabolic diseases
Mental and behavioural disorders
Diseases of the nervous system 3 81 Diseases of the eye and adnexa 2 41 Diseases of the ear and mastoid process
Diseases of the circulatory system 4 177 Diseases of the respiratory system 4 103 Diseases of the digestive system 5 295 Diseases of the skin and subcutaneous tissue
Diseases of the musculoskeletal system and connective tissue
Diseases of the genitourinary system
Pregnancy, childbirth and the puerperium
Congenital malformations, deformations and chromosomal abnormalities
Symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classified
Injury, poisoning and other consequences of external causes
Factors influencing health status and contact with health services
Number of malignant tumour-related diseases n (%) 0b 9 (2) 0�15 1 143 (32) 2 3 (60) 135 (11) 3 1 (20) 44 (7) 4 29 (4) 5 16 (4) >5 1 (20) 40 (10)
Pre-operative antimicrobial prophylaxis n (%)
4 (80) 310 (75) 0�81
Duration of pre-operative antimicrobial prophylaxis (day), median (range)
5 (2–9) 3 (1–11) 0�07
aInternational Statistical Classification of Diseases and Related Health Problems 10th revision (ICD-10) code, including overlap. bPatients whose abnormal tissue samples were found to be not cancerous after post-operative pathological examination. PI, post-operative infection group; NPI, no post-operative infection group. We compared the baseline characteristics of the PI and NPI groups using the Wilcoxon rank sum test for continuous variables, the Yates’ chi-square, or Fisher’s exact test for discrete variables.
Table 4. Antimicrobial therapy for post-operative infection
ATC level5 name na Start of administration (days after ESD) Duration (day)
Sulfamethoxazole and trimethoprim
1 7 2
Metronidazole 1 4 14 Doripenem 1 5 4 Cefazolin 1 2 1 Piperacillin and enzyme inhibitor
1 10 5
Cefmetazole 1 1 4
aIncluding combination therapy. ATC, Anatomical Therapeutic Chemical Classification System.
© 2015 John Wiley & Sons Ltd Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 573–577 576
Post-operative infection of endoscopic T. Muro et al.
general colon surgery. Unlike an open abdominal surgery, ESD does not remove the digestive tract and does not spill bacteria- laden intestinal contents into the abdomen. Supporting this, the present study found that perforation of the digestive tract was an important contributor to post-operative infection. The reduced incidence in spilling bacteria-laden intestinal contents into the abdomen may contribute to the lower post-operative infection rate of ESD. The use of second-generation cephalosporin (cefoxitin or cefotetan) or cefazolin plus metronidazole is recommended as intravenous antimicrobial prophylaxis for colorectal surgery.4,7 In the present study, 76% of pre-operative antimicrobial prophylaxes were second-generation cephalosporin, and the selection of antibiotics used in the study population was shown to be reasonable. On the other hand, the median duration of antimicro- bial prophylaxis was 3�0 (range 1–11) days, which was longer than the reported 24-h duration used in common practice.8 Nelson et al. reported that there was no need for a second intra-operative dose, or any post-operative doses, when the antibiotic was used for the purpose of prophylaxis for ESD or EMR.7 However, Japanese guidelines on the management of infectious diseases have reported that antimicrobial prophylaxis for a duration of 2 days was acceptable, but over 3 days increased the risk of infection with antimicrobial-resistant bacteria.9 Therefore, these results suggested overall adherence to the Japanese guidelines for antimicrobial
prophylaxis duration. Nearly all antimicrobial agents used for post-operative infection were those selective for anaerobic bacteria. Whether these choices in antibiotics were optimal cannot be determined with confidence due to the lack of cultivation test results in the database and is one of the limitations of the present study. Furthermore, the present study was unable to assess treatment result data, such as interviews and palpation informa- tion, nor were images available for investigation. Therefore, post- operative infection was defined in a two-step process, first identifying those patients diagnosed with an infectious disease obtained from the case records and then identifying those who were prescribed an antibiotic.
WHAT IS NEW AND CONCLUSION
Post-operative infection is an exceedingly rare event following ESD-ECN. Pre-operative antimicrobial prophylaxis had no signif- icant effect on post-operative infection following ESD-ECN and thus may be unnecessary. On the other hand, the prevention of digestive tract perforation is important for decreasing the inci- dence of post-operative infection.
1. Toyonaga T, Man-I M, Morita Y et al. The new resources of treatment for early stage colorectal tumors: EMR with small incision and simplified endoscopic submucosal dis- section. Dig Endosc, 2009;21(Suppl 1):S31– S37.
2. Sakamoto N, Osada T, Shibuya T et al. Endoscopic submucosal dissection of large colorectal tumors by using a novel spring- action S-O clip for traction (with video). Gastrointest Endosc, 2009;69:1370–1374.
3. Seitz U, Bohnacker S, SoehendraN. Endoscopic removal of large colon polyps. In: UpToDate, Saltzman JR, eds. Waltham, MA: UpToDate, 2015.
4. Anderson DJ, Sexton DJ. Control measures to prevent surgical site infection following gastrointestinal procedures in adults. In:
Harris A ed. UpToDate. Waltham, MA: UpToDate, 2015.
5. Baum ML, Anish DS, Chalmers TC, Sacks HS, Smith H Jr, Fagerstrom RM. A survey of clinical trials of antibiotic prophylaxis in colon surgery: evidence against further use of no-treatment controls. N Engl J Med, 1981;305:795–799.
6. Min BH, Chang DK, Kim DU, Kim YH, Rhee PL, Kim JJ, Rhee JC. Low frequency of bacteremia after an endoscopic resection for large colorectal tumors in spite of extensive submucosal exposure. Gastrointest Endosc, 2008;68:105–110.
7. Nelson RL, Gladman E, Barbateskovic M. Antimicrobial prophylaxis for colorectal sur- gery. Cochrane Database Syst Rev, 2014;5: CD001181.
8. Bratzler DW, Houck PM, Surgical Infection Prevention Guideline Writers Workgroup. Antimicrobial prophylaxis for surgery: an advisory statement from the National Surgi- cal Infection Prevention Project. Am J Surg, 2005;189:395–404.
9. Mikamo H, Takesue Y, Nakamura H, Fukushima R, Mizusawa M. “Prevention of postoperative infection”. The JAID/JSC Guide to Clinical Management of Infectious Diseases. Guide to Clinical Management of Infectious Diseases committee of The Japa- nese Association for Infectious Diseases and Japanese Society of Chemotherapy ed. The Japanese Association for Infectious Diseases and Japanese Society of Chemotherapy, and Japanese Society of Chemotherapy 2011, p. 182–188.
Table 6. Pre-operative antimicrobial prophylaxis of PI
ATC level5 name n (%) Duration (day), median (range)
Cefotiam 2 (50) 5�0 (5–5) Cefmetazole 1 (25) 2�0 (–) Cefoperazone, combination 1 (25) 9�0 (–)
Table 7. Univariate analyses of PI vs. NPI
PI NPI OR 95% CI
Pre-operative antimicrobial prophylaxis n, (%)
4 (80) 310 (75) 0�73 0�08–6�61
Perforation of the digestive tract n, (%)
1 (20) 6 (1) 17�1 1�66–176�45
PI, post-operative infection group; NPI, no post-operative infection group; OR, odds ratios; 95% CI, 95% confidence interval.
© 2015 John Wiley & Sons Ltd Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 573–577 577
Post-operative infection of endoscopic T. Muro et al.
Copyright of Journal of Clinical Pharmacy & Therapeutics is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder’s express written permission. However, users may print, download, or email articles for individual use.