Artículo Clínico

---

Cross Infection in maxillofacial surgery
Infección quirúrgica en cirugía maxilofacial

 

J. Martínez-Checa¹, V. Fuentes Gómez², E. Jiménez Romano³, M.T. León Espinosa de los Monteros², M. Llamas del Castillo³, M.J. Bermejo Pérez², J. Planes Martínez⁴

---

1 Enfermera.
2 Médico residente.
3 Médico adjunto.
4 Jefe de Servicio.
Servicio de Cirugía Oral y Maxilofacial.
Servicio de Medicina Preventiva.
Hospital Universitario Virgen de las Nieves. Granada. España.

Correspondencia:
J. Martínez Checa
Camino de Ronda 79, 8º A - 18004 Granada. España.

 

Introduction

Nosocomial surgical infection derived from Maxillofacial Surgery has some different connotations than those of other types of surgery. On the one hand, it occurs in territories, such as the mouth, that have a high content of aerobic and anerobic germs. On the other hand, few publications referring to infection after major surgery of the oral cavity and the oropharyngeal region are found in the bibliography.

Our objective is to provide the study result of the infection incidence in the surgical site in the Maxillofacial Surgery Service of the University Hospital Virgen de las Nieves of Granada and to study some risk factors involved in it.

Material and Method

A prospective study of cohorts was performed in the Maxillofacial Surgery Service of the University Hospital Virgen de las Nieves. The study period ranged from September 15, 1999 to November 15, 2000. All the patients hospitalized who underwent surgical interventions in this Service were included. The following patients were excluded from the study: 1) Those cases that were admitted for surgical intervention due to cellulitis after odontogenic infection, regardless of whether it was preceded by dental extraction, 2) cases of patients who underwent dental extractions and who required hospitalization because they had psychic incapacity or because they presented some other abnormality and 3) those cases that suffered exitus in the seven days following the surgical intervention and did not present infection of the surgical site.

The study population was made up of a cohort of 382 patients.

The follow-up was carried out for thirty days after the surgery, this being prolonged to one year in the cases in which an osteosynthesis was performed.

Information was collected with individualized protocols that included the following data: name, age, gender, clinical record number, admission date, main and secondary diagnosis, antibiotic prophylaxis, its type and time, date and type of surgical intervention, contamination degree of the surgical wound, surgical time duration, existence of oral communication, ASA grade of the patient (according to the American Society of Anesthesiologists),¹ main surgeon, surgical reinterventions in an interval that does not surpass one month, date of nosocomial infection and responsible germ, and finally, date of hospital discharge¹.

According to the National Nosocomial Infection Surveillance³ NNIS Study, we classified our patients into four groups:

1. Score 0 . Patients with ASA < 3, surgical intervention time less than two hours and surgical wound contamination grades of the groups clean and clean contaminated.

2. Score 1. Patients with ASA ≥ 3 or surgical intervention time greater than 2 hours or surgical wound contamination grade in the group of contaminated or dirty.

3. Score 2. Patients with two of the risk parameters (ASA superior to 3, time superior to 2 hours or contaminated/ dirty wound).

4. Score 3. Patients with three risk parameters.

The diagnosis of infection was performed following the CDC criteria (1992).²

Statistical analysis

A bivariate analysis to study the association between the possible risk factors and the development of nosocomial infection was performed. The Chi squared test for dichotomic variables, the Student’s t test for the comparison of means between groups and the simple logistic regression for continuous numbers were used. Statistical association was performed in terms of relative risk and 95% confidence interval.

A multivariate analysis was also performed by multiple logistic regresion to identify the predictor factors independent of surgical wound infection.

The database and analysis were performed with the SPSS 10.0 for Windows statistical program.

Results

The total population of the study was 382 patients, 36 of whom presented surgical wound infection, which represents a global incidence of 9.4%.

Study parameters:

1. Age

The mean age of the population was 43.38 years, with a range between 2 and 91 years.

The relationship between age groups and infection is studied in table 1.

If we compare the infection incidence of each age group with the rest of the sample, we obtain a RR of 0.16 (0.05- 0.53), p<0.005 for a group of 20 to 39 years; the RR is 4.17 (2.24-7.76), p<0.005 for the 60-79 year old group. The other age groups did not provide significant data.

2. Gender

A total of 282 surgical interventions were performed in men (73.8%) with 31 infections (11%) and 100 in women (26.2 %) with 5 infections (5%). Although the man/woman infection risk ratio is 2.35 (95% CI: 0.88-6.21), we did not observe statistically significant differences, p= 0.078.

3. Antibiotic prophylaxis

Antibiotic prophylaxis was performed in 323 patients (84.6%). In 231 of them, amoxyclavulanic acid was used at a dose of 1 gram every 6 hours. In 72 cases, clindamycin was used at the dose of 300 mg. every 8 hours and other antibiotics were used in 20 cases (cefuroxime, cloxacillin, vancomycin, etc.). The usage time was not well defined and depended on the criterion of each surgeon. It ranged from 24 hours to seven days in most of the cases.

In 59 cases (15.4%), no antibiotic prophylaxis was performed. These were small surgical interventions, clean, without osteosynthesis that had a short duration.

4. Infection incidence

The patients were divided into two groups: a) traumatological surgery and b) non-traumatological surgery. There were 169 patients (44.2%) in the first group, 3 (1.8%) of whom were infected, while the second group had 213 (55.8%), 33 (15.5%) of whom were infected. The RR for non-traumatological surgery was 8.73 (95% CI: 2.72-27.97). The difference of infection is very significant, with p< 0.005.

5. Study of hospital stay

We have studied two parameters under this point of view, the total days of stay and the days of pre-operative stay. We compared the results between infected and non-infected patients (Table 2).

The mean stay of the infected patients is 3.47 times greater than the non-infected ones, a statistically significant difference (p< 0.005). Delay in the surgical intervention (mean pre-operative stay time) is two times greater in the infected patients than the non-infected ones with a statistically significant difference (p<0.005).

6. Duration of surgical intervention

Mean duration of the surgical intervention is 137.5 minutes, 2.6 times greater in the infected patients than in the non-infected ones with p< 0.005 (Table 3).

To compare the results with the NNIS (National Nosocomial Infection Surveillance) study, T time (75 percentile of time used in the different surgical interventions) was established at 2 hours, although this time was greater in some surgeries (oncologic).

The difference of infection between patients with surgical intervention time less than or equal to two hours and those with a time superior to two hours is statistically significant with a p < 0.005. The RR for the surgical interventions of more than two hours was 7.65 (95% CI: 3.49-17.23) (Table 4).

7. ASA Classification

According to the NNIS study, the patients were divided into two large groups: ASA grade ≤ 2 with a total of 332 (86.9%), 28 of which presented infections (8.4 %) and ASA grade ≥ 3, with 50 (13.1%), 8 of which presented infections (16%). We have not observed a statistically significant association between the ASA grade of the patient and the nosocomial infection (RR: 1.9; 0.91-3.92), p= 0.08.

8. Difference between malignant and benign conditions

We have introduced this parameter, considering that the existence of a malignant condition represents a risk factor due to many circumstances (alteration in immunity, treatment with cytostatics or radiations, etc.). On these bases, the study between benign and malignant conditions produced the following results: benign conditions operated on 243 (63.6 %) with 7 infections (2.9 %) and malignant conditions operated on 139 (36.4%) with 29 infections (20.9 %). The statistical significance is clear with p< 0.005. RR for the patient group with malignant conditions was 7.24 with a 95% CI of 3.26-16.09.

9. Wound contamination degree

According to this parameter, 337 (88.2%) with 31 infections (9.2%) belong to the clean or clean contaminated surgery group and 45 (11.8%) with 5 infections (11.1%) to the contaminated or dirty surgery group. The difference of infection in the bivariate analysis has no statistical significance (p= 0.68), however, if we stratify according to the benign-malignant variable (Cochran test: p<0.05), we find that the variable contamination degree is associated to the infection variable. The positive homogenecity test allows us to accept a risk estimation of 6.08 (95% CI: 1.56-23.73) for the contaminated-dirty surgery group.

Table 5 shows the results regarding the contamination of the surgical field, considering, on the one hand, the oral surgical interventions or those with oral communication, and on the other, those that were not performed in this territory and do not have oral communication.

The RR for the oral/with oral communication interventions was 2.17 with 95% CI of 1.10 – 4.30. The infection difference was statistically significant p= 0.02. If we only select those non-traumatologic and malignant cases, this parameter of surgical interventions performed in the oral cavity/oral communication and the others without oral communication provide the following results (Table 6).

The difference of infection between these two groups is statistically significant, with p= 0.013, and the RR in the group of patients with surgical interventions of malignant conditions of the oral cavity/oral communication is 3.10 (95% CI: 1.14-8.36).

10. Reintervention

Considering this variable, we divided the patients into two groups: those who suffered two or more surgical interventions with intervals not exceeding one month and that represented the value of 27 (7.1%) with 12 infection (44.4%) and those others in which only one surgical intervention was performed, 355 (92.9%) with 24 infections (6.8%). The difference had statistical significance; p < 0.005. The RR for the reintervention group was 6.57 with a 95% CI of 3.71- 11.65.

11. NNIS Index^3,4

If we consider the NISS index as a continuous variable and apply the simple logistic regression, we will obtain an OR for each increase unit of the variable of 3.61 (95% CI: 2.38-5.60) (Table 7).

12. Multivariate analysis

In the multiple logistic regression analysis, all the variables studied were introduced. These included both those that provided significant results in the bivariate analysis as well as those that were considered to have an epidemiologic interest for the study, there being 3 factors that were independent predictive factors of infection: a surgical intervention time over two hours (OR: 7; 3.01-16.25), wound contamination degree (OR: 7.20; 1.25-26.52) and surgical reintervention in a time inferior to one month (OR: 6.29; 2.64- 14.94). The confidence interval width would improve if we had a greater number of cases.

13. Germs responsible

In Table 8, we indicate the germs responsible for the infection. A total of 33.3% was mixed flora (aerobic and anaerobic, representative of the flora that colonizes the oral cavity). The responsible germ was not identified in 11 cases (28.2%), in five of them because no sample was taken for the microbiology study and in the remaining six because there was no bacterial growth. There was purulent exudate in the 11 cases. The total number of germs exceeds by 3 the number of infections because three samples with growth of two germs were identified in the microbiological study.

 

Discussion

Infection incidence in minor surgery such as exodontias is very low, less than 1%,⁵ however, in major surgeries, the infection percentage increases considerably, and can reach values superior to 40%.

The incidence of surgical infection in all the surgical interventions of the Maxillofacial Service of our Center was 9.4%. In the traumatological surgical interventions, the infection percentage was 1.8% and in the non-traumatological ones 15.5%. In the references consulted, we could verify that the infection rates of surgical wound in this type of surgery vary greatly. In cervico-facial carcinogenic surgery, Penel et al.₆ mentioned an infection percentage of 6.6% and Mustafa et al.₇ of 11.6%. According to Penel et al., if this surgery is accompanied by opening of the upper aerodigestive tract mucosae(8), the percentage of infection may reach 41.8% (much greater than that shown in our study); in this setting, as we mentioned previously, the percentage of infection in surgery with oral communication was 12.6%, an intermediate value to that obtained by De Melo in oral cancer⁹ of 22.7% and Magrin et al.¹⁰ in glossectomies with 16 % and to that obtained by Simons et al.¹¹ in head and neck contaminated surgery that showed infection percentages of 8.1%. These percentages are practically the same as those of our study. When we only analyzed the malignant conditions that were performed in the oral cavity or that had communication with it, we obtained an infection incidence of the surgical site of 26.9% while the percentage was 8.7% in those in which there was no communication.

In regards to traumatologic surgery, the infection rates also vary greatly. Abubaker and Rollert¹² mention values of 13.3% of infection in mandible fractures, Iizuka et al.¹³ 6.1% and Heit et al.¹⁴ 3.7%. All of these are superior to the results of our study that were 1.8%.

In the bivariate analysis of the pre-operative risk factors of surgical infection, we found 5 variables with statistical significance: the days of pre-operative stay, duration of the surgical intervention, existence of communication with the oral cavity, existence of reintervention and it being a malignant condition.

Naturally, the causes that make these factors appear as a risk may be diverse. For example, in the case of pre-operative study, the fact that the patient has a general poor condition that may condition an immunodeficiency may be involved. When dealing with a malignant condition, an increase in the duration of the surgical intervention, receiving cytostatic treatment, etc. in addition to deficient immunity may also be involved.

In regards to age, which may be considered as a risk factor, our study provides the following result: if we compare each age group with the rest of the sample, we observe that belonging to the 20 to 39 year old age group would act as a protective factor against the infection while belonging to the 60 to 79 year old group would be a risk factor. As we will see in the following, the age factor loses significance in the multivariate study.

In the bivariate analysis, the wound contamination degree is not statistically significant as a risk factor (unless we stratify it according to the malignancy-benignancy of the condition), however the existence of oral communication is one. Furthermore such an important increase of the infection incidence in the surgery of this territory cannot be explained only by the fact of being clean-contaminated surgery, and other factors that we have not considered in our study (existence of tracheostomy) must play a part.

The physical condition of the patient according to the Classification of the American Society of Anesthesiologists (ASA) is not statistically significant in our study as a risk factor of infection, although this evaluation takes into consideration a series of host factors that are directly related with the risk of infection, including: age, nutritional status, presence of systemic diseases and projected mortality.⁴

In the multivariate analysis, we found three variables as independent predictors of infection: surgical intervention time, surgical wound contamination degree and reintervention. The three parameters are highly verified in all types of surgery and the first two are included as such in the NNIS study.

Finally, in this study, the infection incidence increases step by step as the NISS index score increases.

In regards to the bacteriological analysis and if we consider that 25 (69.4%) of the 36 infections had communication with the oral cavity, it is logical that 33% of our patients would present infection due to mixed flora (aerobic and anaerobic), flora that usually colonizes this territory,¹⁵ and perhaps this percentage would be greater if we had microbiologically identified the 11 infections in our study which could not be identified. Possibly, as we commented on before, being patients receiving antibiotic treatment may interfere in the results of the cultures.

Undoubtedly, it is necessary to unify antibiotic prophylaxis in a future study since, in this one, it has been very diverse, both due to the variety of the antibiotics used as well as by their duration and thus we could not analyze this factor or obtain valid conclusions in this regards. It is also necessary to introduce other risk factors, such as the performance of tracheostomy accompanying the main surgical and the surgeon factor, that could not be included in this study, even though the information had been obtained, as the number of cases assignable to each one of them was not enough to allow for a statistically significant analysis.

Conclusions

Although our rate of infection of the surgical site considered globally cannot be compared with similar studies, since we include such different surgery as facial traumatology, parotid tumors, tongue, cavum, neck tumors, etc., when we break them down, we obtain results that are similar to those mentioned in the references that refer to malignant tumors and that are inferior if we refer to traumatology surgery.

The factors identified in our study as independent predictors of infection were: surgical intervention time, surgical wound contamination grade and reintervention.

References

1. Owens WD, Felts JA, Spitznagel EL. ASA physical satatus Classifications: A Study of Consistency et Rating. Anesthesiology 1978;49:239-43.         [ Links ]

2. Horan TC, Gaynes RP, Martone WJ, Jarvis WR, Emori TG. CDC definitions of nosocomial surgical site infections 1992: A modifications of CDC definitions of surgical wound infections. Infect Control Hos. Epidemiol 1992; 13:606-8        [ Links ]

3. Culver DH, Horan TC, Gaynes RP, et alt. Surgical wound infection rates by wound class, operative procedure and patient risk index. Am J Med (supp B) 1991;3B:152- 7.         [ Links ]

4. Garibaldi RA, Cushing D, Lerer T. Risk Factor for Postoperative Infection. Am J Med (supp B) 1991;3B:158-62.         [ Links ]

5. Peterson LJ. Antibiotic prophylaxis against wound infections in oral and maxillofacial surgery. J Oral Maxillofac Surg 1990;48:617-20.         [ Links ]

6. Penel N, Fournier C, Kara A, Savini J, Lefebre D. Chimiotherapie anti-cancéreuse, facteur de risque d’infections du site opératoire en chirurgie prope cervico- faciale. Presse Med 2001;30:892-5.         [ Links ]

7. Mustafa E, Tashin A. Cefotaxime prophylaxis in mayor non contaminated head and neck surgery: one day versus seven days therapy. J Laryngol Otol 1992;7:30- 2.         [ Links ]

8. Penel N, Lefebre D, Fournier C, Sarini J, Kara A, Lefebre JL. Risk factor for wound infection in head and neck cancer surgery: a prospective study. Head Neck 2001; 23:447-55.         [ Links ]

9. De Melo GM, Ribeiro KC, Kowalski LP, Deheinzelin D. Risk factors for postoperative complications in oral cancer and their prognostic implications. Arch Otolaryngol Head Neck Surg 2001;127:828-33.         [ Links ]

10. Magrin J, Kowalski LP, Saboia M, Saboia RP. Major glossectomy: end results of 106 cases. Eur J Cancer B Oral Oncol 1996;32B:407-12.         [ Links ]

11. Simons JP, Johnson JT, Yu VL, Vickers RM, Gooding WE, Myers EN, Pou AM, Wagner RL, Grandis JR. The role of topical antibiotic prophylaxis in patients undergoing contaminated head and neck surgery with flap reconstruction. Laryngoscope 2001;111:329-35.         [ Links ]

12. Abubaker AO, Rollert MK. Postoperative antibiotic prophylaxis in mandibular fractures: A preliminary randomized, double-blind, and placebo-controlled clinical study. J Oral Maxillofac Surg 2001;59:1415-9.         [ Links ]

13. Iizuka T, Lindqvist C, Hallikainen D, Paukku P. Infection after rigid internal fixation of mandibular fractures: a clinical and radiologic study. J Oral Maxillofac Surg 1991; 49:585-93.         [ Links ]

14. Heit JM, Stevens MR, Jeffords K. Comparison of ceftriaxone with penicillin for antibiotic prophylaxis for compoun mandible fractures. J Oral Maxillofac Surg 1997;83:423-6.         [ Links ]

15. Hart CA. Defence Against Colonization and Infection. En: Vincent JL(ed). Update in Intensive Care and Emergency Medicine, Volume 7. Berlin Heidelberg: Edit. Springer-Verlag 1989:14-5.         [ Links ]