J Infect Chemother (2010) 16:334–339

DOI 10.1007/s10156-010-0108-y

ORIGINAL ARTICLE

Analysis of risk factors for surgical-site infections in 276 oral

cancer surgeries with microvascular free-flap reconstructions
at a single university hospital
Kazunari Karakida • Takayuki Aoki • Yoshihide Ota • Hiroshi Yamazaki •
Mitsunobu Otsuru • Miho Takahashi • Haruo Sakamoto • Muneo Miyasaka

Received: 17 June 2010 / Accepted: 27 July 2010 / Published online: 1 September 2010
Ó Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases 2010

Abstract The purpose of this study was to elucidate the
risk factors for surgical-site infection (SSI) in oral cancer
surgery with microvascular free-flap reconstructions and
to propose appropriate SSI prevention. There were 276
patients who underwent oral cancer surgery with micro-
vascular free-flap reconstructions at the Department of Oral
and Maxillo-facial Surgery of Tokai University Hospital.
The following variables were assessed as risk factors for
SSIs: preoperative variables, including age, sex, body
mass index, American Society of Anesthesiologist’s (ASA)
score, debilitating comorbidities, smoking, alcohol con-
sumption, and Union Internationale Contre le Cancer
Tumor Node Metastasis (UICC-TNM) classification; and
operative variables, including duration of surgery, amount
of blood loss, quantity of blood transfusion, tracheostomy,
area of neck dissection, and previous chemotherapy. Sta-
tistical analysis was conducted to determine whether these
factors constitute risks for SSI. Total overall SSI rate was
40.6% (112/276). When the occurrence of SSI was com-
pared with the variables, ASA score (P = 0.036), T stage
(P = 0.013), duration of surgery (P \ 0.001), blood loss
K. Karakida (&)
H. Sakamoto
Department of Oral and Maxillofacial Surgery,
Tokai University Hachioji Hospital, 1838 Ishikawa-cho,
Hachioji, Tokyo 192-0032, Japan
e-mail: karakida@is.icc.u-tokai.ac.jp
T. Aoki
Y. Ota
H. Yamazaki
M. Otsuru
M. Takahashi
Department of Oral and Maxillofacial Surgery, Tokai University
School of Medicine, 143 Shimokasuya, Isehara,
Kanagawa 259-1193, Japan
M. Miyasaka
Department of Plastic and Reconstructive Surgery,
Tokai University School of Medicine, 143 Shimokasuya,
Isehara, Kanagawa 259-1193, Japan
(P = 0.001), blood transfusion (P = 0.01), and area of
neck dissection (P = 0.009) showed statistical signifi-
cance. Analysis of these variables with a logistic regression
model yielded ASA score and duration of surgery as sig-
nificant factors. There was a tendency for blood loss and
duration of surgery to increase in patients with a high T
stage. A high T stage not only broadens the resection area
and increases surgical invasiveness, it also increases sus-
ceptibility to dead space after microvascular reconstruction
for oral cancer. Particular care in treating the wound should
be taken in surgical patients with high T-stage scores. The
occurrence of SSI is of particular concern in oral cancer
surgery in patients with high ASA scores.
Keywords Surgical-site infection
Microvascular
free-flap reconstructions
Risk factor
Oral cancer
Introduction
The large defect formed by resection of advanced oral
cancer results in severe dysfunction of swallowing and
breathing as well as cosmetic disorder. Recent advances in
microvascular free-flap surgery make it possible to recon-
struct the huge defect according to its complicated mor-
phology. Therefore, surgery is regarded as a standard
treatment for advanced oral cancer. However, duration of
surgery and the amount of blood loss are increased in oral
cancer surgery, there is a greater degree of surgical inva-
sion, which may induce serious postoperative complica-
tions, the most common being surgical-site infection (SSI).
SSI often causes necrosis of the reconstructive flap or
delayed wound healing, deteriorates the patient’s general
condition, and reduces quality of life by delaying recovery.
SSI also increases the length of hospital stay, resulting in

123
J Infect Chemother (2010) 16:334–339
large medical expense. The risk of SSI can also be affected
by medical co morbidities.
Oral cancer surgery is categorized as clean–contami-
nated surgery by the Centers for Disease Control and
Prevention (CDC) [1]. The frequency of SSI in oral cancer
surgery is high compared with other clean–contaminated
types of surgery. Possible reasons for this are as follows:
(1) After resection of advanced oral cancer, the oral cavity
and the cervical region communicate. (2) Cervical tissue is
exposed to saliva contaminated with bacteria and respira-
tory-tract secretions. (3) In ablative defects with complex
morphology, dead space might form after wound closing.
However, there are few evidence-based studies on pre-
venting SSI following microvascular free-flap reconstruc-
tion, and little is known about the possible mechanisms of
SSI in oral cancer surgery. To prevent SSI and improve
treatment safety, clarifying the risk factors would be
valuable.
We retrospectively studied 276 patients who underwent
oral cancer surgery with microvascular free-flap recon-
structions at a single university hospital to elucidate the
risk factors for SSI.
Patients and methods
Patient characteristics
We examined 276 patients at Tokai University Hospital
from 1996 to 2005. Patients who had overt infection around
the tumor before surgery and those who had previous
radiotherapy were excluded. Tumors were staged using the
guidelines established by the Union Internationale Contre
le Cancer Tumor Node Metastasis (UICC-TNM) classifi-
cation that included clinical findings as well as computed
tomography (CT), magnetic resonance (MR) tomography,
ultrasonography, and bone scintigraphy. Preoperative body
weight and height were measured, and body mass index
(BMI) values were calculated. Debilitating comorbidities,
smoking, alcohol consumption, and American Society of
Anesthesiologists (ASA) preoperative assessment scores
were recorded.
Surgical regimen
For all patients, primary-site resection was combined with
neck dissection, and the oral and cervical wound commu-
nicated. For the defect that was produced after resection,
immediate reconstruction using a microvascular free flap
was performed by a plastic surgeon. The suture used for the
oral mucosa and reconstructive flap was 3-0 polyglactin
910 braided thread (3-0 VicrylÒ). For the cervical wound,
4-0 polydioxanone monofilament thread (PDSIIÒ) was
335
used for the subcutaneous layer and 4-0 nylon monofila-
ment thread for the cutaneous layer. Several low-pressure
suction drains (4 mm J-VACÒ drain) were inserted to
prevent subcutaneous hematoma, exudates, and dead space.
Regimen of preoperative therapy
Some patients received neoadjuvant chemotherapy using
peplomycin sulfate, cisplatin, and fluorouracil. Surgery was
performed after recovery from side effects of neoadjuvant
chemotherapy, such as myelosuppression. Ampicillin or
cefmetazole was administered for antimicrobial chemo-
prophylaxis during surgery. The intervals of antibiotic
administration were once an hour before surgery, once
during surgery, and twice a day for 3 days after surgery, at
a dose of 1 g each time. The low-pressure suction drains
removed 30 ml or less of waste fluid. Sutures were
removed 7–10 days after surgery. Postoperative radiother-
apy was delivered to the neck in patients with stage pN2b
disease. If a residual tumor was identified at the surgical
margins of the resected sections, the primary site was
treated with postoperative radiotherapy. However, postop-
erative radiotherapy was performed after confirming the
absence of SSI, and there were no patients who developed
SSI during radiotherapy.
Definition of SSI
CDC criteria for SSI was adopted and slightly modified to
include infections that developed up to 30 days after sur-
gery, including both wound incision and organ/space
infection. Infections at sites remote from the wound,
including respiratory infection, urinary tract infection, and
catheter infection, were considered remote infections.
Bacterial isolation
Organisms were isolated from an aseptically obtained
culture of fluid or tissue from the superficial incision.
Risk factors for SSIs
To assess for SSI, we included only patients who devel-
oped SSI within the first 30 days after surgery. Only the
most severe episode of SSI was included for patients who
developed more than one SSI. The following variables
were assessed as risk factors: Preoperative variables
included sex, age, BMI, ASA score, debilitating comor-
bidities, smoking, alcohol consumption, UICC-TNM clas-
sification, and previous chemotherapy. Operative variables
included duration of surgery, amount of blood loss, quan-
tity of blood transfusion, tracheostomy, and area of neck
dissection.
123
336 J Infect Chemother (2010) 16:334–339

Table 1 Primary tumor site Table 3 Relationship between patient characteristics and risk of
surgical-site infection (SSI) (categorical data)
Site No. of patients Surgical-site
(n = 276) infection (%) Factor Categories No. of patients SSI P value
(n = 276) (%)
Tongue 84 25 (29.8)
Lower gingiva 83 42 (50.6) Gender Male 170 69 (40.6) 0.874
Buccal mucosa 32 15 (46.9) Female 106 42 (39.6)
Floor of mouth 32 13 (40.6) ASA score 1 163 56 (34.4)
Upper gingiva 21 9 (42.9) 2 109 54 (49.5) 0.036
Palate 9 4 (44.4) 3 4 1 (25.0)
Intraosseous of the mandible 5 4 (80) Debilitating Yes 162 69 (42.6) 0.468
comorbidities No 110 42 (38.2)
Maxillary sinus 5 1 (20)
Parotid gland 2 0 (0) Smoking Yes 141 57 (40.4) 0.943
Lower lip 2 0 (0) No 135 54 (40.0)
Masseter muscle 1 0 (0) Alcohol Yes 136 53 (39.0) 0.677
No 140 58 (41.4)
T stage 1 18 7 (38.9) 0.013
Table 2 Pathological diagnosis 2 123 39 (31.7)
3 29 9 (31.0)
Diagnosis No. of patients Surgical-site
(n = 276) infection (%) 4a 95 51 (53.7)
4b 7 4 (57.1)
Squamous cell carcinoma 255 103 (40.4)
N stage 0 145 61 (42.1) 0.891
Verrucous carcinoma 8 4 (50)
1 43 16 (37.2)
Salivary gland tumors 7 1 (14.3)
2a 2 1 (50)
Malignant non-epithelial neoplasms 4 2 (50)
2b 63 23 (36.5)
Ameloblastic carcinoma 2 1 (50)
2c 17 7 (41.2)
3 3 2 (66.7)
Statistical analysis Blood transfusion Yes 111 55 (49.5) 0.01
No 165 56 (33.9)
The chi-square (v2) test (for categorical data) and the Tracheostomy Yes 190 79 (41.6) 0.493
Student’s t test (for continuous data) were performed to No 86 32 (37.2)
assess the relationship between potential risk factors and Neck dissection Unilateral 191 67 (35.1) 0.009
outcome of SSI. Odds ratios (ORs) and 95% confidence Bilateral 85 44 (51.8)
intervals (CIs) were computed for associations. When more Previous Yes 139 55 (39.6) 0.825
than two categories were analyzed, a global chi-square, chemotherapy No 137 56 (40.9)
associated P value, and ORs comparing the category with ASA American Society of Anesthesiologists
lower risk were computed. A logistic regression model was P \ 0.05
used to evaluate variables found to be associated with SSIs
by univariate analysis. Analyses were performed using whom reconstruction was performed, T stage was mostly
SPSS version 16.0 (SPSS Inc., Chicago, IL, USA). A T2 or higher, with 258 cases. N stage was N0 in 145
P value \0.05 was considered statistically significant. cases, N2b in 63 cases, and N1 in 43 cases (Table 3).
ASA score was commonly 1, with 163 cases, and was 2
in 109 cases and 3 in four cases (Table 3). There were
Results 162 patients with debilitating comorbidities, comprising
59.6% of the total (Table 3). One hundred and forty-one
Patient characteristics patients were smokers, and 136 regularly drank alcohol,
comprising about half of the total (Table 3). SSI occurred
The most common primary tumor sight among the 276 in 112 (40.6%) of the 276 patients.
patients was the tongue, in 84 cases (Table 1). The
pathological diagnosis in the majority of cases was Surgery
squamous cell carcinoma (Table 2). Patients consisted of
169 men and 107 women with an average age of Neck dissection was performed in all cases: unilater-
63 (range 23–91) years. Because these were patients for ally in 191 patients and bilaterally in 85 (Table 3).

123
J Infect Chemother (2010) 16:334–339 337

Microvascular free flap was used for reconstruction in all patients with and without SSI showed a significant differ-
cases, with rectus abdominis musculocutaneous flap used ence (P = 0.036) (Table 3). When disease characteristics
in 221 cases, radial forearm flap in 26, anterior thigh flap of patients with and without SSI were compared, TNM
in 13, fibular osteoseptocutaneous flap in seven, and stages showed significant difference only in the T stage
plural flap in nine. In addition, a titanium reconstruction (P = 0.013) (Table 3).
plate was used in 48 patients who had mandible resection In a comparison of operation characteristics, patients
(Table 4). with SSI showed significantly longer surgery duration
(P \ 0.001), blood loss (P = 0.001), quantity of blood
Bacteriology transfused (P = 0.01), and neck dissection area
(P = 0.009) than did patients without SSI. That is, as
The main bacteria detected in patients with SSI were surgical invasiveness increased, the frequency of SSI also
methicillin-resistant Staphylococcus aureus (MRSA) in increased (Tables 3, 5). Remote infection occurred in 50
12, Pseudomonas aeruginosa in six, Klebsiella pneumo- cases (18.1%). These infections were respiratory in 23
niae in three, Serratia marcescens in two, S. epidermidis cases, catheter in 23, urinary tract in six, and multiple sites
in three, Enterobacter cloacae in two, Enterobacter aero- in two. Statistical analysis of the relationship between SSI
genes in two, Enterococcus spp in two, and others in and remote infection showed that the occurrence of SSI
eight. was significantly increased in patients who developed
remote infections (P \ 0.001). Univariate analysis showed
Risk factors for SSIs that the significant risk factors for SSI were ASA score, T
stage, duration of surgery, amount of blood loss, quantity
When characteristics of patients with and without SSI of blood transfusion, and area of neck dissection. These
were compared, age and sex showed no statistical signif- factors were analyzed with a logistic regression model.
icance (SSI vs. age: P = 0.278; SSI vs. sex P = 0.874) Multivariate analysis yielded the ASA score and duration
(Tables 3 and 5). Statistical analysis of ASA scores of of surgery as significant risk factors. ORs for ASA score
and duration of surgery were 1.771 and 1.004, respectively
Table 4 Reconstruction free-flap material (Table 6).
Free flap No. of patients Surgical-site
(n = 276) infection (%)
Discussion
Radial forearm flap 26 8 (30.8)
Rectus abdominis 221 90 (40.7) Oral and maxillofacial surgery is classified in the surgical
musculocutaneous flap wound classification of the CDC’s guideline as class II
Anterolateral thigh flap 13 3 (23.1) (clean–contaminated), the same as surgery for the ali-
Fibular osteoseptocutaneous flap 7 4 (57.1) mentary canal. The incidence of SSI in class II surgeries in
Plural free flapsa 9 6 (66.7) regions such as the gastroenterological tract is reportedly
Total 276 5.8–22.4% [2–4]. Among oral and maxillofacial surgeries,
Any flap with reconstruction plate 48/276 25 (52.1) oral cancer surgery with microvascular free-flap recon-
a
Bilateral rectus abdominis musculocutaneous flap 9 4, latissimus struction has the greatest invasiveness and a high incidence
dorsi myocutaneous flap ? scapular bone flap 9 2, bilateral rectus of SSI: 36.4 –50% [5–9]. A so-called pull-through resec-
abdominis musculocutaneous flap ? deltopectoral flap 9 1, Rectus tion is often conducted in this operation, which allows the
abdominis musculocutaneous flap ? anterolateral thigh flap 9 1,
rectus abdominis musculocutaneous flap ? fibular osteoseptocutane-
cervical tissue and free flaps to be exposed to saliva and
ous flap 9 1 alimentary-tract or respiratory-tract secretions, which may

Table 5 Relationship between patient characteristics and risk of surgical-site infection (SSI) (continuous data)
Factor Overall average Average with SSI Average without SSI P value

Age (years old) 62.2 ± 12.39 61.2 ± 11.78 62.8 ± 12.77 0.278
Body mass index 22.2 ± 3.44 22.2 ± 3.64 22.2 ± 3.30 0.905
Duration of surgery (min) 523.2 ± 123.91 558.5 ± 126.61 498.8 ± 116.29 \0.001
Blood loss (ml) 762.8 ± 500.08 879.4 ± 554.80 682.9 ± 443.17 0.001
± Standard deviation
P \ 0.05

123
338
Table 6 Multivariate logistic regression model for developing sur-
gical-site infection (SSI)
Factor P value Odds ratios (95% CI)
ASA score 0.023 1.771 (1.083–2.896)
Duration of surgery 0.015 1.004 (1.002–1.007)
CI confidence interval, ASA American Society of Anesthesiologists
P \ 0.05
increase the risk of SSI. To curb the incidence of SSI, we
investigated the risk factors for SSI in three categories:
patient characteristics, disease characteristics, and opera-
tion characteristics. According to the CDC guideline,
patient characteristics that are factors for increasing the risk
of SSI are age, nutritional status, and diabetes. In this
study, univariate analysis showed that among factors of
patient characteristics, patients with a high ASA score have
a high incidence of SSI. However, no significant relation-
ship between SSI and sex, age, BMI, debilitating comor-
bidities, smoking, or alcohol consumtion was observed. It
seems likely that the ASA score showed a significant dif-
ference as an index for the comprehensive assessment of
debilitating comorbidities, age, obesity, etc. In disease
characteristics, T stage showed a statistically significant
difference, and the incidence of SSI increased with
increasing size of the invasion area of the tumor. The
resection area naturally becomes larger with a higher T
stage. A larger resection area increases surgical stress, and
the formation of dead space also needs to be considered.
However, N stage showed no statistically significant dif-
ference in the development of SSI.
Our research also made clear that the development of
remote infections is very likely in patients who develop
SSI. The reason for this is possibly that the performance
status of a patient who does not receive appropriate treat-
ment for SSI deteriorates, and the patient becomes sus-
ceptible to remote infections as a compromised host.
Therefore, early detection and prompt treatment of SSI is
considered one of the most important factors of perioper-
ative care. However, if a remote infection develops first
and is not treated appropriately, the patient becomes a
compromised host, which is considered a factor leading to
SSI.
The bacteria isolated in SSI are nosocomial infection
pathogens, such as MRSA, P. aeruginosa, K. pneumoniae,
or other enteric bacteria. The term of gram-negative aero-
bic bacteria in postoperative infection is undetermined.
Weber described the importance of gram-negative bacteria
in postoperative infection [10]. On the other hand, Johnson
stated that the isolated gram-negative bacteria were
more likely due to colonization than infection [11]. Gram-
negative bacteria and MRSA were mostly resistant to
123
J Infect Chemother (2010) 16:334–339
ampicillin and cefmetazole, although these antibiotics were
appropriate for indigenous oral microflora, which may
cause colonization of new ampicillin-resistant strains from
the environment. The use of antibiotics, including duration
and treatment intervals, should be considered further.
Multivariate analysis showed the significant risk factors
for SSI to be ASA score and duration of surgery. This may
indicate that the risk of SSI is higher when there are
debilitating comorbidities and the duration of surgery is
long with extensive excision. The National Nosocomial
Infection Surveillance System (NNIS) regards surgical
wound classification, duration of surgery, and ASA score
as the three most important risk factors for SSI [1]. In our
analysis, the duration of surgery and ASA score were sta-
tistically significant, in accordance with the NNIS findings.
Oral cancer surgery with microvascular free-flap recon-
struction can leave dead space or lead to ruptured sutures.
Oral organs are constantly moving in association with
deglutition, mastication, and the cough reflex, and it is
difficult to keep the wound still. This leads to delayed
wound healing and is possibly influential in the develop-
ment of SSI. It is also reported that the infection rate is
lower when active rather than passive drainage is used to
drain pooled blood and exudates from the wound [12].
From these results, improving techniques and surgical
skills to shorten the duration of surgery and reduce blood
loss are important. Techniques to eliminate dead space and
the use of appropriate drainage are also important for
preventing SSI.
References
1. Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis WR.
Guideline for prevention of surgical site infection, 1999. Hospital
Infection Control Practices Advisory Committee. Infect Control
Hosp Epidemiol. 1999;20:250–78.
2. Lilani SP, Jangale N, Chowdhary A, Daver GB. Surgical site
infection in clean and clean-contaminated cases. Indian J Med
Microbiol. 2005;23:249–52.
3. Chattopadhyay R, Zaroukian S, Potvin E. Surgical site infection
rates at the Pontiac Health Care Centre, a rural community hos-
pital. Can J Rural Med. 2006;11:41–8.
4. Uchino M, Ikeuchi H, Tsuchida T, Nakajima K, Tomita N,
Takesue Y. Surgical site infection following surgery for inflam-
matory bowel disease in patients with clean-contaminated
wounds. World J Surg. 2009;33:1042–8.
5. Penel N, Fournier C, Lefebvre D, Lefebvre JL. Multivariate
analysis of risk factors for wound infection in head and neck
squamous cell carcinoma surgery with opening of mucosa. Study
of 260 surgical procedures. Oral Oncol. 2005;41:294–303.
6. Lotfi CJ, Cavalcanti Rde C, Costa e Silva AM, Latorre Mdo R,
Ribeiro Kde C, Carvalho AL, et al. Risk factors for surgical-site
infections in head and neck cancer surgery. Otolaryngol Head
Neck Surg. 2008;138:74–80.
7. Penel N, Mallet Y, Roussel-Delvallez M, Lefebvre JL,
Yazdanpanah Y. Factors determining length of the postoperative
J Infect Chemother (2010) 16:334–339
hospital stay after major head and neck cancer surgery. Oral
Oncol. 2008;44:555–62.
8. Liu SA, Wong YK, Poon CK, Wang CC, Wang CP, Tung KC.
Risk factors for wound infection after surgery in primary oral
cavity cancer patients. Laryngoscope. 2007;117:166–71.
9. Ogihara H, Takeuchi K, Majima Y. Risk factors of postoperative
infection in head and neck surgery. Auris Nasus Larynx. 2009;36:
457–60.
10. Weber RS, Raad I, Frankenthaler R, Hankins P, Byers RM,
Guillamondegui O, et al. Ampicillin-sulbactam vs clindamycin in
339
head and neck oncologic surgery. The need for gram-negative
coverage. Arch Otolaryngol Head Neck Surg. 1992;118:1159–63.
11. Johnson JT, Kachman K, Wagner RL, Myers EN. Comparison of
ampicillin/sulbactam versus clindamycin in the prevention of
infection in patients undergoing head and neck surgery. Head
Neck. 1997;19:367–71.
12. Batstone MD, Lowe D, Shaw RJ, Brown JS, Vaughan ED,
Rogers SN. Passive versus active drainage following neck dis-
section: a non-randomised prospective study. Eur Arch Otorhi-
nolaryngol. 2009;266:121–4.
123