Abstract
Superficial surgical site infections (SSIs) are one of the most common postoperative complications of hepatectomy for liver cancer. The objective of this study is to clarify the risk factors and determine a clinical prediction score for SSIs after partial hepatectomy for malignant tumour. A total of 812 consecutive patients were enrolled who underwent partial hepatectomy for liver malignant tumour from January 2017 to December 2017. Univariate and multivariate analyses were conducted to identify the risk factors for SSIs. Clinical prediction score was then constructed using coefficients of identified significant predictors. Risk stratification was then carried out by receiver operating characteristic curve analysis. Of all the 812 patients, SSIs were observed in 31 (3.82%) patients. A multivariate analysis identified four predictors as independent risk factors for SSIs, which were splenomegaly, perioperative blood transfusion, intensive care unit (ICU) admission, and low postoperative serum albumin concentration (<35 g/L). Clinical prediction score ranged from 0 to 4.6 with its discrimination concordance (C) statistic of 0.70 (95% confidence interval [CI] 0.59, 0.81). Risk stratification classified these patients into low, moderate, and high risk in SSIs. This risk score system may credibly stratify the risk of SSIs with relatively high sensitivity and specificity. Splenomegaly, history of blood transfusion, ICU admission, and postoperative serum albumin concentration less than 35 g/L could be used to predict SSIs with acceptable discrimination. This clinical risk score system may be useful in prediction of SSIs after hepatectomy for malignant tumours.
Keywords: albumin level, blood transfusion, hepatectomy, splenomegaly, surgical site infections
1. INTRODUCTION
Superficial surgical site infections (SSIs) are one of the most common hospital‐acquired infections among hepatectomy patients, with a significant impact on the length of hospital stay, medical cost, and the morbidity and mortality rate.1, 2 Thus, accurate prediction of SSIs is essential in clinical practice, which may be helpful for risk stratification and adopting effective prevention and treatment strategies promptly.
The incidence of SSIs after hepatectomy has been reported to be 3.1% to 25.2%.3, 4, 5, 6, 7, 8, 9, 10 Unlike those with metastatic liver tumours, patients with primary liver cancer often have chronic liver disease that may be susceptible to SSIs.11 Studies have shown that cirrhosis and low serum albumin level were the independent risk factors for SSIs in the hepatocellular carcinoma (HCC) patients after hepatectomy.10, 12, 13 The incidence of primary liver cancer, especially HCC, varies greatly in different regions of the world, among which more than half of the HCC cases worldwide occur in China.14 However, there is no valid clinical prediction score system for SSIs after hepatectomy for malignant tumour. Thus, there is an urgent need to establish a valid clinical prediction score system for SSIs after hepatectomy for primary liver cancer patients in China, which may also possibly be applied worldwide.
2. MATERIALS AND METHODS
We retrospectively enrolled 823 consecutive patients who underwent hepatectomy for HCC from January 2017 to December 2017 in the First Affiliated Hospital of Zhejiang University, China. A total of 11 patients were excluded from this study for the following reasons: 2 patients had concomitant gastrointestinal resection, 3 patients underwent radiofrequency ablation combined with liver surgery, and 6 patients' wound status was not properly documented. The remaining 812 patients were analysed, and their characteristics are summarised in Table 1. SSIs were defined as infection either with incisional involvement of the skin alone or with the involvement of deep tissues or organs.15 SSIs were assessed before discharging home, at 1‐week, 2‐week, and 1‐month follow‐up.
Table 1.
Risk factors for SSI in the patients with primary liver cancer after radical resection
| Risk factors | SSI (number %) | No SSI (number %) | P value |
|---|---|---|---|
| Patients related | |||
| Gender | .689 | ||
| Male | 25 (3.08) | 606 (74.63) | |
| Female | 6 (0.74) | 175 (21.55) | |
| Age number | .237 | ||
| ≥60 | 18 (2.22) | 368 (45.44) | |
| <60 | 13 (1.6) | 413 (50.74) | |
| Smoking | .468 | ||
| Yes | 6 (0.74) | 196 (24.14) | |
| No | 25 (3.08) | 585 (72.04) | |
| History of alcohol | .87 | ||
| Yes | 10 (1.23) | 263 (32.39) | |
| No | 21 (2.59) | 518 (63.79) | |
| Diabetes | .229 | ||
| Yes | 6 (0.74) | 84 (10.34) | |
| No | 25 (3.08) | 697 (85.84) | |
| Hypertension | .499 | ||
| Yes | 10 (1.23) | 209 (25.74) | |
| No | 21 (2.59) | 572 (70.44) | |
| ASA classification | .881 | ||
| I‐II | 10 (1.23) | 539 (66.38) | |
| III‐IV | 21 (2.59) | 242 (29.80) | |
| NNIS classification | .743 | ||
| 0–1 | 25 (3.08) | 660 (81.28) | |
| 2 | 6 (0.74) | 121 (14.90) | |
| History of hepatitis B | .062 | ||
| Yes | 14 (1.73) | 483 (59.48) | |
| No | 17 (2.09) | 298 (36.70) | |
| Cirrhosis | .031 | ||
| Yes | 17 (2.09) | 280 (34.48) | |
| No | 14 (17.24) | 501 (61.70) | |
| Splenomegaly | .001 | ||
| Yes | 12 (1.48) | 127 (15.64) | |
| No | 19 (2.34) | 744 (91.64) | |
| Albumin (preoperative) | 1.000 | ||
| >35 | 30 (3.70) | 758 (93.35) | |
| ≤35 | 1 (0.12) | 23 (2.83) | |
| Albumin (postoperative) | .002 | ||
| >35 | 5 (0.62) | 342 (42.12) | |
| ≤35 | 26 (3.20) | 439 (54.06) | |
| With other infections | .771 | ||
| Yes | 3 (0.37) | 52 (6.40) | |
| No | 28 (3.45) | 729 (89.78) | |
| Operation related | |||
| Operative type | .017 | ||
| Laparoscope | 2 (0.25) | 197 (24.26) | |
| Laparotomy | 29 (3.57) | 584 (71.92) | |
| Operative time (minutes) | .000 | ||
| ≥240 | 20 (2.46) | 250 (30.79) | |
| <240 | 11 (1.35) | 531 (65.39) | |
| Blood loose (mL) | .325 | ||
| ≥200 | 15 (1.85) | 309 (38.05) | |
| <200 | 16 (1.97) | 472 (58.13) | |
| Blood transfusion | .012 | ||
| Yes | 5 (0.62) | 35 (4.31) | |
| No | 26 (3.20) | 702 (86.45) | |
| ICU admission | .056 | ||
| Yes | 7 (0.86) | 79 (9.73) | |
| No | 24 (2.96) | 702 (86.45) | |
| TACE before operative | .472 | ||
| Yes | 6 (0.74) | 103 (12.78) | |
| No | 25 (3.08) | 678 (83.50) | |
Abbreviations: ASA, American Society of Anaesthesiologists; ICU, intensive care unit; NNIS, National Nosocomial Infections Surveillance; TACE, Transcatheter Arterial Chemoembolization.
Baseline characteristics of the patients were described. Predictors were presented as number (%) and were analysed using the χ² test and Fisher exact test as appropriate. A multivariate analysis was performed by logistic regression to identify independent parameters correlating with SSIs. Variables with P values <.10 in the univariate analysis were included in the multivariate analysis with the backward method in the case of exclusion of some potential prognostic factors. Using the coefficients of the simplified prediction model, a risk score system for SSIs was then created. The optimal cut‐off value was determined using the receiver operating characteristic (ROC) curves. **Odds ratios (ORs) and their 95% confidence intervals (CIs) are reported. The following 20 variables were examined as potential risk factors: gender, age, history of smoking and alcohol taking, diabetes, hypertension, the American Society of Anaesthesiologists classification, the National Nosocomial Infections Surveillance classification, history of hepatitis B virus infection, liver cirrhosis, splenomegaly, serum albumin level (preoperative and postoperative), other organ infections, operative type, operative time span, blood loss, blood transfusion, intensive care unit (ICU) admission, and transcatheter arterial chemoembolization (TACE) before operation. A two‐sided P value of <.05 was considered statistically significant. All these analyses were performed by using SPSS 19.0.
For this retrospective study, the protocol was approved by an independent ethics committee of The First Affiliated Hospital, College of Medicine, Zhejiang University.
3. RESULTS
3.1. The characteristics of clinical data
After excluding 11 patients for reasons mentioned above, a total of 812 patients were available for analysis. Nearly all operations were elective (97.9%). Average age was 57.73 (21‐84) years old, and 74% of patients were male. In this study, 31 patients developed SSIs during follow‐up, with the incidence of 3.82%, among which 28 patients and three patients presented incisional infection and deep tissue infection, respectively. All the patients with SSIs were treated non‐operatively successfully.
3.2. Risk factors for SSIs
Univariate logistic regression analysis was performed to assess association between each risk factor and SSIs. The results were shown in Table 1. For easy application, continuous variables were categorised into two groups. Among 20 predictors (14 patients related and 6 surgery related), 6 factors were statistically significant which was defined as P < .05. These risk factors were cirrhosis (P = .031), splenomegaly (P = .001), postoperative serum albumin level less than 35 g/L (P = .002), operative type (P = .017), operative time (P = .000) and blood loss (P = .012), which were associated with an increased risk of SSIs. The six parameters shown by univariate analysis vary significantly between patients with and without SSIs, which were thus further analysed by multivariate logistic regression (Table 2). This analysis identified four predictors as independent risk factor for SSIs, including splenomegaly (P = .000), blood transfusion (P = .043), ICU admission (P = .036), and the level of albumin (postoperative) (P = .023) with the OR of 4.82, 3.09, 2.71, and 3.19, respectively.
Table 2.
Risk factors for SSI: a multiple logistic regression
| Variables | Coefficient | SE | P value | OR | Score |
|---|---|---|---|---|---|
| Splenomegaly | 1.574 | 0.343 | .000 | 4.827 | |
| Yes | 1.6 | ||||
| No | 0 | ||||
| Blood transfusion | 1.128 | 0.558 | .043 | 3.09 | |
| Yes | 1.0 | ||||
| No | 0 | ||||
| ICU admission | 1.000 | 0.478 | .036 | 2.718 | |
| Yes | 1.0 | ||||
| No | 0 | ||||
| Albumin (postoperative) | 1.162 | 0.511 | .023 | 3.195 | |
| ≤35 | 1.0 | ||||
| >35 | 0 |
Abbreviations: OR, odds ratio; SSI, superficial surgical site infection.
3.3. Clinical prediction score for SSIs
Using the coefficients of significant predictors, a risk score system for SSIs was then created (Table 3). The risk prediction equation was written as
 |
Table 3.
Risk stratification of prediction values of scoring system
| Risk classification | Scores | SSI (%) | Sensitivity (%) | Specificity (%) |
|---|---|---|---|---|
| Low | <1 | 4 (12.9%) | 87.1 | 30.3 |
| Moderate | 1‐2 | 11 (35.5%) | 51.6 | 75.8 |
| High | >2 | 16 (51.6%) | 45.2 | 87.6 |
Abbreviations: OR, odds ratio; SSI, superficial surgical site infection.
The C statistics of this model was 0.70 (95% CI 0.59, 0.81) indicating acceptable discrimination of SSIs from non‐SSIs. The risk score ranged from 0 to 4.6, and the optimal cut‐off value was determined using the ROC curves. The clinical prediction score classified patients into low risk (score <1, sensitivity 87.1%, and specificity 30.3%), moderate risk (score 1‐2, sensitivity 51.6%, and specificity 75.8%), and high risk (score >2, sensitivity 45.2%, and specificity 87.6%) groups for SSIs. This risk score system accurately stratifies the risk of post‐hepatectomy SSIs. Patients with a risk score <1 were less likely to be SSIs (12.9%), while a score >2 was associated with a high incidence of SSIs (51.6%).
4. DISCUSSION
In this study, 812 patients were included, and the total incidence of SSIs was 3.82%, which was comparable to the rate of SSIs reported elsewhere previously. Among 20 potential predictors (14 patients related and 6 surgery related), 4 independent risk factors for SSIs were identified finally, including splenomegaly, blood transfusion, ICU admission, and the low postoperative serum albumin level (<35 g/L). Previous studies considered elderly patients and the presence of diabetes mellitus16 and long operative time11 as independent risk factors for SSIs. In contrast, these variables were not identified as independent risk factors for SSIs in our study. Importantly, the predictive risk factors for SSIs we discovered in this study including splenomegaly and the low postoperative serum albumin level are all associated with decompensated liver function, which indicated that HCC patients with hepatic dysfunction are more susceptible to SSIs.
Then, by using the coefficients of the simplified prediction model, a risk score system for SSIs was established. The C statistics of this model was 0.70, indicating acceptable discrimination of SSIs from non‐SSIs. The risk score ranged from 0 to 4.6, and the optimal cut‐off value was determined using the ROC curves. Patients with a risk score >2 were associated with a high likelihood of SSIs (51.6%), which showed that risk stratification was easy to apply by just counting the variables occurring at the end of the operation.
5. CONCLUSIONS
In conclusion, this study has shown that splenomegaly, blood transfusion, ICU admission, and the postoperative albumin level less than 35 g/L are independent risk factors for SSIs. The clinical prediction score system for SSIs has been developed, which may be useful in prediction of SSIs after hepatectomy in HCC patients. However, external validation and the impact of this model on clinical practice still need to be assessed in the future.
CONFLICT OF INTEREST
The authors declare no potential conflict of interest.
AUTHOR CONTRIBUTIONS
S.S.Z. was responsible for the conception. J.W.S. and Z.W.N. wrote the manuscript. Y.G.Q., N.Z.W., and B.W. generated the research question, checked, and analysed the data. All authors participated in the drafting of the manuscript and approved the final version.
ACKNOWLEDGEMENTS
This study was supported by the Research on Public Welfare Technology and Social Development Project of Zhejiang Provincial Bureau of Science and Technology (2017C33069).
Shen J, Ni Z, Qian Y, Wang B, Zheng S. Clinical prediction score for superficial surgical site infections: Real‐life data from a retrospective single‐centre analysis of 812 hepatectomies. Int Wound J. 2020;17:16–20. 10.1111/iwj.13209
Juwei Shen and Zuowei Ni contributed equally to this study.
Funding information Research on Public Welfare Technology and Social Development Project of Zhejiang Provincial Bureau of Science and Technology, Grant/Award Number: 2017C33069
REFERENCES
- 1. Garwood RA, Sawyer RG, Thompson L, Adams RB. Infectious complications after hepatic resection. Am Surg. 2004;70:787. [PubMed] [Google Scholar]
- 2. Imamura H, Seyama Y, Kokudo N, et al. One thousand fifty‐six hepatectomies without mortality in 8 years. Arch Surg. 2003;138:1198‐1206. [DOI] [PubMed] [Google Scholar]
- 3. Wu CC, Yeh DC, Lin MC, Liu TJ, P'eng FK. Prospective randomized trial of systemic antibiotics in patients undergoing liver resection. Br J Surg. 1998;85:489‐493. [DOI] [PubMed] [Google Scholar]
- 4. Virani S, Michaelson JS, Hutter MM, et al. Morbidity and mortality after liver resection: results of the patient safety in surgery study. J Am Coll Surg. 2007;204:1284‐1292. [DOI] [PubMed] [Google Scholar]
- 5. Togo S, Matsuo K, Tanaka K, et al. Perioperative infection control and its effectiveness in hepatectomy. J Gastroenterol Hepatol. 2007;22:1942‐1948. [DOI] [PubMed] [Google Scholar]
- 6. Moreno Elola‐Olaso A, Davenport DL, Hundley JC, Daily MF, Gedaly R. Predictors of surgical site infection after liver resection: a multicentre analysis using National Surgical Quality Improvement Program data. HPB(Oxford). 2012;14:136‐141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Shiba H, Ishii Y, Ishida Y, et al. Assessment of blood‐products use as predictor of pulmonary complications and surgical infection after hepatectomy for hepatocellular carcinoma. J Hepatobiliary Pancreat Surg. 2009;16:69‐74. [DOI] [PubMed] [Google Scholar]
- 8. Okabayashi T, Nishimori I, Yamashita K, et al. Risk factors and predictors for surgical site infection after hepatic resection. J Hosp Infect. 2009;73:47‐53. [DOI] [PubMed] [Google Scholar]
- 9. Harimoto N, Shirabe K, Abe T, et al. Prospective randomized controlled trial investigating the type of sutures used during hepatectomy. World J Gastroenterol. 2011;17:2338‐2342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Kobayashi S, Gotohda N, Nakagohri T, Takahashi S, Konishi M, Kinoshita T. Risk factors of surgical site infection after hepatectomy for liver cancers. World J Surg. 2009;33:312‐317. [DOI] [PubMed] [Google Scholar]
- 11. Sadamori H, Yagi T, Shinoura S, et al. Risk factors for organ/space surgical site infection after hepatectomy for hepatocellular carcinoma in 359 recent cases. J Hepatobiliary Pancreat Sci. 2013;20:186‐196. [DOI] [PubMed] [Google Scholar]
- 12. Pessaux P, Msika S, Atalla D, Hay JM, Flamant Y. French Association for Surgical ResearchRisk factors for postoperative infectious complications in noncolorectal abdominal surgery: a multivariate analysis based on a prospective multicenter study of 4718 patients. Arch Surg. 2003;138:314‐324. [DOI] [PubMed] [Google Scholar]
- 13. Mansour A, Watson W, Shayani V, Pickleman J. Abdominal operations in patients with cirrhosis: still a major surgical challenge. Surgery. 1997;122:730‐735. [DOI] [PubMed] [Google Scholar]
- 14. Liu W, Liu Q, Huang Q, et al. Time trend analysis of primary liver cancer incidence in Sihui county of Guangdong Province, China (1987‐2011). BMC Cancer. 2016. Oct 12;16(1):796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. CDC NNIS System . National Nosocomial Infections Surveillance (NNIS) system report, data summary from January 1992 to June 2004, issued October 2004. Am J Infect Control. 2004;32:470‐485. [DOI] [PubMed] [Google Scholar]
- 16. Huo TI, Lui WY, Huang YH, et al. Diabetes mellitus is a risk factor for hepatic decompensation in patients with hepatocellular carcinoma undergoing resection: a longitudinal study. Am J Gastroenterol. 2003;98:2293 e2298. [DOI] [PubMed] [Google Scholar]