Volume 86, Issue 9 p. 1097-1102
Free Access

Post-cesarean surgical site infections according to CDC standards: rates and risk factors. A prospective cohort study

HANS KRISTIAN OPØIEN

Corresponding Author

HANS KRISTIAN OPØIEN

Department of Obstetrics and Gynaecology, Asker og Bærum Sykehus, HF, Norway

Department of Microbiology, Asker og Bærum Sykehus, HF, Norway

: Hans Kristian Opøien, Department of Obstetrics and Gynecology, Rikshospitalet-Radiumhospitalet Medical Center, 0027, Oslo, Norway. E-mail: [email protected]Search for more papers by this author
ANNELILL VALBØ

ANNELILL VALBØ

Department of Obstetrics and Gynaecology, Asker og Bærum Sykehus, HF, Norway

Department of Microbiology, Asker og Bærum Sykehus, HF, Norway

Search for more papers by this author
ANETTE GRINDE-ANDERSEN

ANETTE GRINDE-ANDERSEN

Department of Obstetrics and Gynaecology, Asker og Bærum Sykehus, HF, Norway

Department of Microbiology, Asker og Bærum Sykehus, HF, Norway

Search for more papers by this author
METTE WALBERG

METTE WALBERG

Department of Obstetrics and Gynaecology, Asker og Bærum Sykehus, HF, Norway

Department of Microbiology, Asker og Bærum Sykehus, HF, Norway

Search for more papers by this author
First published: 31 December 2010
Citations: 109

Abstract

Objectives. The aim of this study was to document the true incidence of post-cesarean surgical site infections (SSI), according to the definition of the US Centers for Disease Control and Prevention (CDC), and to identify independent risk factors for infection. Design. Prospective population-based cohort study in Norway. Setting. Sykehuset Asker og Bærum HF, a secondary community hospital, associated with the University of Oslo (UiO), Norway, accounting for 2,000 deliveries per year. Participants. All cesarean deliveries during a 12-month period from September 2003. Main outcome measures. Rate and risk factors for SSI. Results. The total rate of SSI was 8.9%, with an observation period of 30 days post-operatively, compared to 1.8% registered at hospital discharge. The total response rate was 100%. There was no significant difference in SSI rate in elective or emergency cesarean section (CS), respectively. All SSI were superficial. We found 2 significant independent risk factors: operating time ≥38 min and body mass index (BMI) >30. Conclusion. The rate of SSI is underestimated if the observation time is limited to the hospital stay. Operating time exceeding 38 min substantially increases the risk of SSI. The finding of no significant difference in SSI rate between elective and emergency CS should lead to a different approach concerning the use of antibiotics: subgroup at risk (operating time ≥38 min and BMI >30) may benefit from antibiotics in relation to the operation, whether the CS is an emergency or elective operation.

Introduction

Cesarean section (CS) is one of the most commonly performed surgical procedures in hospitals. In Norway, as in other industrialised western countries, the incidence has increased during the last 30 years, from 2% in 1968, to 12.6% in 1990, to 15.8% in 2002 (1–5). Additionally, a considerable variation has been registered between different obstetric departments (6–20%) (6). In 1998, the Norwegian Medical Association invited all maternity units to participate in a ‘Breakthrough Project’ on cesarean delivery, in order to accelerate improvements of the medical practice by cycles of setting aims, establishing measures, and developing and testing changes (7). In one of the publications emerging from this project, an unexpected high overall complication rate during hospital stay was reported (21.4%). Wound infection affected 2.2% of the patients having regional anesthesia (4). The incidence of maternal post-operative complications varies in the literature (4), (8–19). However, the average length of post-operative hospital stay has decreased in recent years, leaving certain maternal post-operative complications, such as surgical site infections (SSI), not revealed at time of hospital discharge. As many of the patients later consult primary health care services in Norway, the incidence of SSI might be underestimated, when SSI is defined, according to the US Centers for Disease Control and Preventions (CDC), as an infection occurring within 30 days after a surgical procedure (without any implant) (20) (Figure 1).

Details are in the caption following the image

Criteria for defining surgical site infection (SSI), from US Centers for Disease Control and Prevention.

The aim of this study was to document the true incidence of post-cesarean SSI according to the CDC definition, and to identify independent risk factors for infection.

Material and methods

Data were collected from all women who underwent cesarean delivery in Asker and Bærum Hospital during a 12-month period from September 2003. Nine women were excluded because of linguistic communication problems. The study group then consisted of 326 women. The Asker and Bærum Hospital is a community hospital, serving 150,000 people, having approximately 2,000 deliveries per year, and a CS rate of 13.8% in the study period.

The CS was defined as elective if the procedure was performed at least 8 h after the decision was made to operate. CS performed within 8 h of the decision, were classified as an emergency operation.

All women having an emergency CS received prophylactic antibiotics; Cefalotin (2 g) as an intravenous single dose after the umbilical cord was clamped. Women presumed allergic to Cefalotin were given Clindamycin (600 mg) instead. Women with elective CS received no antibiotic prophylaxis. This procedure is in accordance with the Norwegian national guidelines (21).

All surgical interventions were performed jointly by an attending physician and a resident. Our standard operative procedure, after cutaneous disinfection with chlorhexidin, is a low transverse incision followed by a cross-sectional incision on the lower uterine segment. The abdomen is closed with a continuous single layer suturation (Polyglactine 910®, absorbable) of the uterotomy and the fascia, and an intracutaneous suture of the skin.

On the day of discharge, the surgical wound was inspected and classified according to CDC definitions. Two of the authors (HKO and AGA) were responsible for this procedure. All women were meticulously instructed about signs and symptoms of SSI, and were asked to contact the hospital if they suspected an infection, within 30 days after the operation. A questionnaire to record the wound healing during the 30 days observational period was handed out. Every contact with the health care system and every conceivable mode of treatment of the wound were registered. The questionnaire was to be completed and returned in a prepaid envelope 30 days after the surgical intervention. Non-responders received a reminder. Those still not responding were traced and interviewed by telephone. All 326 women were registered without any drop-outs.

Statistical analyses were performed by means of Statistical Packages for Social Sciences (SPSS, version 14). Comparisons between groups were obtained by χ2 test, t-test and Mann–Whitney test. Stratified analyses and logistic regression analyses were used to identify risk factors for infection. Results are reported with 95% confidence interval (CI).

Results

In total, 29 out of 326 women (8.9%) developed an SSI after CS. Demographic data and pre-operative variables were registered on a form. These variables are listed in Table I, together with data from the post-operative hospital stay collected from the patient record.

Table I. Pre- and post-operative characteristics of the study group.
Emergency CS (n = 201) Elective CS (n = 125) p-value
Age, mean (range) 31.8 (21–42) 33.4 (24–41) 0.001
Nulliparous (%) 124 (61.7) 40 (32.0) <0.01
BMI, mean (range) 29.5 (18–50) 28.1 (20–42) 0.013
No. of women with intact/ruptured membranes before the CS 57 (intact membranes) 144 (ruptured membranes) 122 (intact membranes) 3 (ruptured membranes)
Time from rupture of membranes to CS Mean:19.4 h (range: 1–105) Mean: 0 h
No. of women with temperature > 38°C before the CS (%) 21 (10.4) 0 <0.001
No. of women receiving antibiotic treatment before the CS (%) 17 (8.5) 4 (3.2) 0.066
No. of women receiving prophylactic antibiotics in relation to CS (%) 178 (88.6) 5 (4) <0.001
No. of women with bacterial vaginosis confirmed pre-op. (%) 1 (0.5) 1 (0.7) 1.00
No. of women with vaginal Streptococcus-B confirmed pre-op. (%) 7 (5) 1 (0.8) 1.00
No. of women on current immunosuppressive treatment (%) 0 2 (1.5%) 0.16
No. of women on current anti-viral treatment (%) 1 (0.5%) 0 1.00
Number of post-op. days in maternity ward, mean (range) 5.1 (1–19) 4.6 (1–10) 0.01
No. of women with temperature >38°C post-op. 13 (6.5) 4 (3.2) 0.43
No. of women having SSI diagnosed before hospital discharge (%) 4 (2) 2 (1.6) 1.00

Among the 326 women undergoing CS in the study period, 125 (38.5%) had an elective CS and 201 (61.5%) had an emergency CS. The SSI rate in the elective and emergency groups were 9 (7.2%) and 20 (10%), respectively. The difference in SSI between the 2 groups of CS was not significant (p = 0.24). In the emergency CS group, 11 women (3.4%) had their surgical procedures performed with a maximum of urgency, and did not receive preoperative cutaneous disinfection. None of these 11 women developed SSI. Out of the 29 women with SSI, 23 recognised the wound infection after hospital discharge, accounting for 79.3% of the SSI. The 23 women were examined by a physician at the hospital. All infections were superficial, and the description of the treatment is listed in Table II.

Table II. Treatment of post-cesarean SSI.
Total no. of SSI (all superficial) No treatment Local treatment (ointment containing antibiotics) Systemic antibiotics Surgical revision Systemic antibiotics + surgery
29 3 5 8 4 9

Variables expected to be related to SSI were analysed in relation to SSI: age, nationality, parity, body mass index (BMI), bacterial vaginosis during pregnancy, elapsed time from rupture of membranes, duration of labour, maternal fever during labor, emergency nature of the operation, duration of the operation, volume of blood loss, antibiotic prophylaxis (or antibiotic treatment of co morbidity), and length of hospital stay. Variables with a p-value < 0.10 were included simultaneously as independent variables in a logistic regression analysis. This analysis revealed two significant, independent risk factors for post-cesarean SSI.

  • 1

    Operating time ≥ 38 min; p = 0.026, unadjusted OR = 2.5 (95% CI: 1.1–5.2), adjusted OR = 2.4 (95% CI: 1.1–5.3). The group was divided into quartiles according to operation time, and within each quartile the risk of infection was calculated. We then found that in the 3 first quartiles, the risk of infection was similar, whilst in the fourth quartile, it was increased. The cut of time between the fourth and other quartiles was 38 min.

  • 2

    BMI > 30; p = 0.007, with an adjusted OR = 2.8 (95% CI: 1.3–6.2).

A cross tabulation (Table III) of the two risk factors and SSI showed a high risk group of 31 women with BMI > 30 and operating time ≥ 38 min, in which as many as 9 (29%) developed SSI, indicating an elevated risk for infection in this group.

Table III. A cross tabulation of the two risk factors (BMI > 30 and operation time ≥ 38 min) and SSI, demonstrating a high risk group of 31 women.
BMI ≤ 30 BMI > 30
Operation time < 38 min 5.5% (9/164) 9.6% (7/73)
Operation time ≥ 38 min 7.3% (4/55) 29% (9/13)

Discussion

We found a post-cesarean SSI rate of 8.9% when the observational period was 30 days according to the CDC standard. The study represents an entire geographical cohort of women requiring CS for various clinical and socio-psychological reasons. We believe the SSI incidence in the present study to be the true incidence, due primarily to the 100% response in the follow-up, and to the CDC criteria of 30 days observation time. The strict criteria for the diagnosis of the SSI to be either superficial or profound were followed, and only two of the participating obstetricians were involved in the classification procedure.

The possible pitfall of misjudging a wound infection when the woman herself is reporting it has been tentatively controlled for by investigating all her contacts with the health care system in the observation period. The patient's assessment of the wound has low positive predictive value, but more important, a negative predictive value as high as 98% according to Withby et al., arguing that very few cases of SSI are missed when women assess the wound status (22). The size of the study population may represent a limitation for estimating moderately increased risk factors. The documented increased risk of prolonged operation time ≥ 38 min was, however, statistically well supported.

Post-cesarean infection rate has been reported to vary from 1.1 to 25% in different studies, but comparison is difficult because of different definitions, classifications and observation time (9), (13), (14), (16–18), (23–25). Few studies report SSI registrations from elective and emergency cesareans separately, or apply the CDC criteria. In the report from the Norwegian ‘Breakthrough Project’, the incidence of wound infection was as low as 2.2 and 2.7% in the group of women with regional and general anesthesia, respectively (4). This is in accordance with earlier reports, such as Ehrenkranz et al. reporting a wound infection rate of 1.1%, and Nielsen et al. reporting an infection rate of 1.6% (9), (24). But in both reports, the observation time was confined to the hospital stay. However, Couto et al. and Tran et al. found 9.6 and 9.8%, respectively, SSI when the follow-up period was 30 days after hospital discharge (17), (18).

In a Cochrane Review, only 14 of the 66 randomised studies included are from western European countries, and among them, only 4 are from the Nordic countries (26). Prophylactic use of antibiotics for all CS interventions had been the general recommendation in this database of systematic reviews. The main argument for this was prevention of SSI and endometritis. Endometritis is argued to occur as often as in 61% of all CS, and is reduced by up to 75% when prophylactic antibiotics are applied (26). In two different Norwegian studies, the incidence of endometritis was reported to be < 1% (27), (28). This supports the view that SSI is the more important post-cesarean infection in Norway, and hence our main consideration concerning indications for the use of prophylactic antibiotics.

In the present study, no significant difference in infection rate was detected between elective and emergency operations, but, emergency CS has been demonstrated to be a major risk factor for complications in several previous studies (12), (24), (29–32). According to our national consensus, prophylactic antibiotics are recommended in all emergency CS (21). If our use of prophylactic antibiotics has had the anticipated effect, a real SSI difference between elective and emergency operations may have been masked. The use of prophylactic antibiotics in relation to elective CS has been reported to vary between the different obstetric departments in Norway (6).

Bagratee et al. demonstrated a marginal and non-significant reduction in SSI rate (0.8%) with the use of prophylactic cefoxitin in elective CS, and Ruiz et al. found no statistically significant differences in elective CS in ‘infection related complications’ (11), (33).

Our study demonstrates that an operating time ≥ 38 min and BMI > 30 are independent risk factors for post-cesarean SSI. The association between increasing BMI (or subcutaneous tissue thickness), and the safety of surgical procedures and post-operative infection has been thoroughly documented (10), (15), (17), (24), (34), (35). But, to our knowledge, longer lasting operations with a specific cut of time, independent of the peri-operative blood loss, has not been described as an operative risk factor for SSI. The need for an experienced obstetrician to be present and responsible for the operative procedure is supported.

The implication of this study is that the true incidence of post-operative SSI is underestimated if post-discharge surveillance is not implemented. Actual recommendations on this issue vary between different countries. We believe that women with significant increase in risk factors (operating time exceeding 38 min and BMI > 30) should receive antibiotics, regardless of whether the CS is an elective or an emergency operation. As the use of antibiotics has individual and environmental side-effects that must to be taken into account, limitation of their use has long been advocated (36). The benefit of expanded use of prophylactic antibiotics in order to reduce the incidence of SSI should be questioned, when treating only the subgroups at increased risk is a justified option.

Conclusion

The rate of SSI is underestimated if the observation time is limited to the hospital stay. Operating time exceeding 38 min together with a BMI > 30 substantially increased the risk of SSI in the present study. To prevent an expanded use of antibiotics, and to further prevent SSI, subgroups at risk should be identified and treated with antibiotics in relation to the operation.

Acknowledgement

We thank Professor Per Bergsjø for valuable comments on the manuscript and Professor Leiv Sandvik for assistance with the statistics.