Association of the Duration of Antibiotic Therapy With Major Surgical Site Infection in Cochlear Implantation

Key Points

Question

Is there a difference in major infections after cochlear implantation among patients treated with prolonged antibiotic treatment or treated with only single-dose antibiotic prophylaxis?

Findings

Among 1180 patients in this cohort study, postoperative cochlear implant infection was rare, was less common among those who received prolonged antibiotic treatment vs antibiotic prophylaxis, and was less likely to occur in adults than in children.

Meaning

Prolonged antibiotic treatment is associated with a reduction in postoperative cochlear implant infection risk compared with antibiotic prophylaxis.

This cohort study describes the rates of major surgical site infection for patients undergoing cochlear implantation who receive prolonged antibiotic treatment compared with those who receive a single perioperative dose of antibiotic prophylaxis.

Abstract

Importance

Infection after cochlear implantation is a rare but serious event that can lead to meningitis. There is no consensus on prevention of infection in these patients, and each center applies its own strategy.

Objective

To describe the rates of major surgical site infection for patients undergoing cochlear implantation who receive prolonged antibiotic treatment compared with those who receive a single perioperative dose of antibiotic prophylaxis.

Design, Setting, and Participants

Retrospective cohort study of patients who underwent cochlear implantation between January 1, 2011, and July 8, 2015, with a postoperative follow-up of 1 to 3 years. In this multicenter study at 8 French university centers, 1180 patients (509 children and 671 adults) who underwent cochlear implantation during this period were included.

Interventions

Prolonged antibiotic treatment vs single-dose antibiotic prophylaxis.

Main Outcomes and Measures

Major infection and explantation.

Results

Among 1180 patients (509 children [51.7% female] with a mean [SD] age of 4.6 [3.8] years and 671 adults [54.9% female] with a mean [SD] age of 54.8 [17.0] years), 12 patients (1.0%) developed a major infection, with 4 infections occurring in the prolonged antibiotic treatment group and 8 infections occurring in the antibiotic prophylaxis group (odds ratio, 2.45; 95% CI, 0.73-8.17). Children (9 of 509 [1.8%]) were more likely to develop infection than adults (3 of 671 [0.4%]). Among children, 4 infections occurred in the prolonged antibiotic group (n = 344), and 5 infections occurred in the antibiotic prophylaxis group (n = 158) (odds ratio, 2.78; 95% CI, 0.74-10.49). Among adults, 3 infections occurred in the antibiotic prophylaxis group (n = 365), whereas no infections occurred in the prolonged antibiotic treatment group (n = 290).

Conclusions and Relevance

After cochlear implantation, infection was rare, was less common among those who received prolonged antibiotic treatment, and was less likely to occur in adults than in children.

Introduction

Cochlear implantation is a method of hearing rehabilitation for people with severe to profound sensorineural hearing loss. The procedure is well standardized and safe, as demonstrated by the low rate of surgical complications reported to date after 30 years of use of this hearing rehabilitation technique.¹ That said, complications can occur, and one of the most common and worrisome after cochlear implant surgery is infection. Surgical site infection (SSI) has many repercussions. These SSIs may be minor and respond to medication, or they may cause significant morbidity and necessitate reoperation and sometimes explantation. Cochlear implantation SSI may not only jeopardize hearing rehabilitation but also endanger the patient’s life by causing meningitis^2,3 or other complications involving the brain or meninges. Over time, the surgical approaches have been made as minimally invasive as possible, with the size of retroauricular incisions similar to that of tympanoplasty.^4,5,6 The length of the procedure is also much shorter than when the technique was first introduced, rarely lasting more than 2 hours.^7,8 For many reasons, most teams prefer a minimally invasive surgical procedure often referred to as “soft surgery.”^9,10,11 This type of surgical technique aims to preserve remaining hearing function and natural cochlear structure, first by limiting the cochleostomy and ensuring that closure is complete, and second by limiting forces applied inside the cochlea. For example, the natural access route of the round window limits cochlear damage as much as possible, thereby reducing the risk of perioperative infection of the perilymph and cerebrospinal fluid and consequently the risk of meningitis.^12,13 Furthermore, because no drilling is required in most round window insertion, intracochlear fibrous tissue formation can be limited.

By reducing the size of the skin incision, limiting the cochlear opening and scar tissue formation, and decreasing the overall length of the procedure, soft surgical techniques are likely to lower the risk of postoperative infection. Questions remain about the use of antibiotics. Their effectiveness for preventing postoperative infection has not been demonstrated, and there is no consensus on the subject. France is one of the largest consumers of antibiotics in Europe¹⁴ (30% higher than the European average), yet the bacterial disease rate is not lower than in other European countries. Because antibiotic use is known to cause bacterial resistance with dramatic consequences,¹⁵ it is reasonable to ask whether antibiotics should be used and what form treatment should take: should we choose prolonged antibiotic treatment or only single-dose antibiotic prophylaxis?

In an attempt to answer this question, we retrospectively collected postoperative data from 8 French centers and focused on infectious complications occurring during 4 years, with at least 1 year of postoperative follow-up. Our main objective was to obtain a general picture during this 4-year period regarding the incidence of infectious complications in patients undergoing cochlear implantation to investigate whether a particular method of using antibiotics clearly stood out, particularly whether antibiotic prophylaxis sufficed compared with standard, prolonged antibiotic treatment.

Methods

Study Design and Population

We conducted a retrospective cohort study at 8 French university centers with a cochlear implant program that agreed to take part in the trial. Our investigations included all patients who underwent unilateral or bilateral cochlear implant surgery between January 1, 2011, and July 8, 2015, with a follow-up of 1 to 3 years. The statistical analysis was performed in June 2017. During this period, most cochlear implants were still in use. Recently archived patient files were easily accessible, with almost no missing data. The team responsible for the cochlear implantation program at each center took charge of collecting data about their own patients while maintaining anonymity. Therefore, it was possible to collect the following epidemiological data: sex, age, date of surgery, brand of implant, antibiotic regimen (prolonged antibiotic treatment or antibiotic prophylaxis), comorbidities, mode of administration of antibiotics, occurrence of infection, and, if infection occurred, the time after surgery, bacterium responsible, and occurrence of meningitis, as well as the need for antibiotic treatment or explantation. The data were drawn from surgical reports, discharge prescriptions, and anesthesia files. All institutional review boards at the 8 centers approved this study, and the need for patient informed consent was waived owing to the use of deidentified data.

The primary analysis was of an exploratory nature and aimed to describe current practice at 8 cochlear implantation centers regarding prevention of postoperative infection. A further analysis then compared the infection rate in patients who received more than 48 hours of antibiotics (prolonged antibiotic group) vs that in patients who received 48 hours or less of antibiotics (antibiotic prophylaxis group). The infection rate was also investigated in terms of age, with a comparison of children (younger than 16 years) and adults (16 years or older). Patients 16 years or older were considered adults as in most of the studies published on this subject. To enhance data anonymity, the participating cities were not named but were identified by number.

It appeared that the prescription of prolonged antibiotic treatment or antibiotic prophylaxis depended on the preferences at each center. One center always prescribed prolonged antibiotic treatment. A second center prescribed prolonged antibiotic treatment only for children. Three other centers always prescribed antibiotic prophylaxis. At a sixth center, only adults were prescribed antibiotic prophylaxis. At the remaining 2 centers, the prescription was less systematically offered. In actuality, it depended on the discretion of the surgeon or the anesthesiologist. Therefore, clear consensus on this topic is missing.

Definition of SSI

According to the French National Authority for Health, SSI is considered to be treatment related if the infection occurred within 30 days of an intervention or within 1 year of the intervention if an implant, prosthesis, or prosthetic material is placed. In our study, we investigated the occurrence of infection up to 1 year after the date of implantation. We recorded cases of major SSI that required hospital treatment ranging from intravenous antibiotics to surgical drainage or explantation.

Surgical Technique

The surgical technique differed little from one center to the next. A C-shape or occasionally an S-shape retroauricular incision was made, usually concave toward the front, extending superiorly and posteriorly. A musculoperiosteal flap was then created to permit mastoid drilling. In most cases, the electrode array was inserted using posterior tympanotomy. The cochlear implant was placed in the temporal region under the musculoperiosteal flap after a bony well was created, if desired, but most often was simply secured in a subperiosteal pocket, sometimes with the help of transmuscular sutures or self-tapping screws. The electrode array was inserted through the round window or via a transpromontorial cochleostomy, which was closed using a musculoaponeurotic or temporal fascia graft after the electrode array was inserted. Last, the retroauricular wound was closed in layers.

Statistical Analysis

Statistical analyses were conducted using a software program (Stata, version 13; StataCorp LP). Continuous data were expressed as means (SDs) according to the statistical distribution (normality was assessed by means of the Shapiro-Wilk test). The statistical tests were 2 sided, with a significance level of .05. Comparisons between independent groups were performed using the χ² test or Fisher exact test for categorical variables. For quantitative variables, the t test was used, or the Mann-Whitney test was used if the t test assumptions were not met (ie, normality and homoscedasticity as studied using the Fisher-Snedecor test). When appropriate, a generalized linear model (logistic regression for a dichotomous end point) was used. The results were expressed as odds ratios and 95% CIs.

Results

Global Findings

Altogether, 1180 patients received implants during the study period, of whom 509 were children (51.7% female; mean [SD] age, 4.6 [3.8] years) and 671 were adults (54.9% female; mean [SD] age, 54.8 [17.0] years). Of the 1180 patients, 23 (1.9%) received neither prolonged antibiotic treatment nor antibiotic prophylaxis, 634 (53.7%) received prolonged antibiotic treatment for 3 days or more, and 523 (44.3%) received antibiotic prophylaxis. In those patients who received prolonged antibiotic treatment, 568 of 634 (89.6%) took antibiotics for 7 days; for the others, the shortest treatment lasted 3 days in 2 patients, and the longest treatment lasted 21 days in 2 patients. Of the 523 patients who received antibiotic prophylaxis (<3 days), only 1 patient received treatment up to day 1 and 6 patients up to day 2, whereas all the other 516 patients received a single prophylactic dose.

In the 1157 patients who received either prolonged antibiotic treatment or antibiotic prophylaxis, there were 12 major postoperative infections (Table 1). Five of the infections were in female patients, and 7 were in male patients.

Table 1. Characteristics of 1157 Patients Who Received Either Prolonged Antibiotic Treatment or Antibiotic Prophylaxis.

Variable	All (N = 1157)		Children (n = 502)		Adults (n = 655)
	No Infection (n = 1145)	Infection (n = 12)	No Infection (n = 493)	Infection (n = 9)	No Infection (n = 652)	Infection (n = 3)
Sex, No.
F	593	5	240	3	353	2
M	552	7	253	6	299	1
Age, median (IQR), y	32 (4-59)	4 (2-30)	3 (2-6)	2 (1-4)	57 (43-68)	58 (49-71)
City, No. (%)
1	83 (7.2)	1 (8.3)	24 (4.9)	1 (11.1)	59 (9.0)	0
2	31 (2.7)	1 (8.3)	10 (2.0)	1 (11.1)	21 (3.2)	0
3	468 (40.9)	2 (16.7)	266 (54.0)	2 (22.2)	202 (31.0)	0
4	43 (3.8)	0	18 (3.7)	0	25 (3.8)	0
5	46 (4.0)	1 (8.3)	17 (3.4)	1 (11.1)	29 (4.4)	0
6	143 (12.5)	1 (8.3)	49 (9.9)	1 (11.1)	94 (14.4)	0
7	115 (10.0)	0	48 (9.7)	0	67 (10.3)	0
8	216 (18.9)	6 (50.0)	61 (12.4)	3 (33.3)	155 (23.8)	3 (100)
Brand of implant, No. (%)
Advanced Bionics	93 (8.1)	1 (8.3)	41 (8.3)	1 (11.1)	52 (8.0)	0
Cochlear	509 (44.5)	5 (41.7)	294 (59.6)	3 (33.3)	215 (33.0)	2 (66.7)
MED-EL	213 (18.6)	4 (33.3)	83 (16.8)	3 (33.3)	130 (19.9)	1 (33.3)
Neurelec	330 (28.8)	2 (16.7)	75 (15.2)	2 (22.2)	255 (39.1)	0
Antibiotic regimen, No. (%)
Prolonged antibiotic treatment	630 (55.0)	4 (33.3)	340 (69.0)	4 (44.4)	290 (44.5)	0
Antibiotic prophylaxis	515 (45.0)	8 (66.7)	153 (31.0)	5 (55.6)	362 (55.5)	3 (100)

Abbreviation: IQR, interquartile range.

In the prolonged antibiotic treatment group (n = 634), 3 infections occurred after treatment had been conducted for 7 days, and 1 infection occurred after 10 days, totaling 4 infections (0.6%). In the antibiotic prophylaxis group (n = 523), 8 infections (1.5%) were reported for a difference of 0.9% (95% CI, −0.3% to 2.1%) and an odds ratio of 2.45 (95% CI, 0.73-8.17). Children (9 of 509 [1.8%]) were more likely to develop infection than adults (3 of 671 [0.4%]).

Five infections occurred early during the first month after surgery. Seven infections occurred between 1 month and 1 year after surgery. No cases of meningitis were observed. Comorbidities were unremarkable in 8 of the 12 infection cases. There was 1 previous case of otogenic meningitis that caused the hearing loss, 1 case of chronic otitis with cholesteatoma that was surgically treated on the side opposite the implant, 1 case of metabolic disease (mitochondrial cytopathy), and 1 case of eczema accompanied by asthma.

Findings in Children

There were 502 children who received antibiotics, among whom 344 (68.5%) had prolonged antibiotic treatment. Nine infections occurred in the pediatric population, 4 of 344 (1.2%) in the prolonged antibiotic treatment group and 5 of 158 (3.2%) in the antibiotic prophylaxis group, for an absolute risk difference of 2.0% (95% CI, −0.8% to 6.5%) and an odds ratio of 2.78 (95% CI, 0.74-10.49). In 5 of these 9 cases of infection, the cochlear implant was not removed. In 2 patients, the infectious pattern began as acute otitis media; in 4 patients, the infectious pattern began as an SSI. Precise data concerning the presentation of infection in the remaining 3 patients was not available (Table 2).

Table 2. Infection Cases in Children.

Patient No./Sex/Age, y	Prolonged Antibiotic Treatment	Antibiotic Prophylaxis	Time After Surgery	Bacterium Responsible	Occurrence of Meningitis	Implant Removal	Comorbidities	Clinical Presentation of Infection
1/M/1.4	Augmentin 80 mg/kg/d for 10 d	No	8 d	Enterobacter cloacae	No	No	No	NA
2/F/4.0	No	Augmentin 360 mg	3 wk	Staphylococcus aureus	No	No	Mitochondrial cytopathy	Acute otitis media
3/M/1.3	Augmentin 80 mg/kg/d for 7 d	No	12 mo	Negative	No	No	No	SSI
4/M/1.7	No	Cefazolin 250 mg	4.5 mo	S aureus plus Streptococcus constellatus	No	Yes	No	Acute otitis media
5/F/3.2	No	Cefazolin 1 g	8 mo	Multiresistant Staphylococcus epidermidis	No	Yes	No	SSI
6/F/1.3	Augmentin 80 mg/kg/d for 7 d	No	5 mo	Negative	No	Yes	Growth failure	NA
7/M/11.0	Augmentin 1.5 g/d for 7 d	No	7 d	Negative	No	No	Food allergies, eczema	NA
8/F/1.5	No	Augmentin 80 mg/kg	1 mo	Haemophilus influenzae	No	No	No	SSI
9/M/4.0	No	Cloxacillin sodium 100 mg/kg	8 mo	S aureus	No	Yes	Otogenic meningitis	SSI

Abbreviations: NA, not available; SSI, surgical site infection.

Findings in Adults

There were 655 adults, among whom 365 (55.7%) had antibiotic prophylaxis. Analysis of the adult population showed that 3 infections occurred in the antibiotic prophylaxis group, whereas none occurred in the prolonged antibiotic treatment group (0% [0 of 290] vs 0.8% [3 of 365]). Two of the 3 had an SSI, whereas the last one had external otitis. The cochlear implant was removed in 1 patient with SSI (Table 3).

Table 3. Infection Cases in Adults.

Patient No./Sex/Age, y	Prolonged Antibiotic Treatment	Antibiotic Prophylaxis	Time After Surgery	Bacterium Responsible	Occurrence of Meningitis	Implant Removal	Comorbidities	Clinical Presentation of Infection
1/F/58.4	No	Cefazolin 2 g	2 mo	Staphylococcus aureus plus Peptostreptococcus	No	No	No	SSI
2/F/70.9	No	Augmentin 3 g	8 mo	Pseudomonas aeruginosa	No	Yes	Surgically treated congenital cholesteatoma	SSI
3/M/49.3	No	Cefazolin 2 g	18 d	Escherichia coli plus P aeruginosa	No	No	No	External otitis

Abbreviation: SSI, surgical site infection.

Other Findings

Overall, there were 75 sequential cochlear implantations during the study period. Only 1 of them, in a child, developed infection, but the implant was eventually salvaged (at center 1).

Among 1157 patients, the most common brand of implants used was Cochlear (514 [44.4%]), followed by Neurelec (now Oticon Medical) (332 [28.7%]), MED-EL (217 [18.8%]), and Advanced Bionics (94 [8.1%]). Table 1 summarizes the overall results, in particular the main brands of implants used and the number of infections associated with each brand. The infection rate varied from 1.0% (5 of 514) for Cochlear, 0.6% (2 of 332) for Neurelec, 1.8% (4 of 217) for MED-EL, and 1.1% (1 of 94) for Advanced Bionics. The infection rate of approximately 1% did not differ significantly across the 4 brands. The infection rate across the different city locations of the 8 centers varied from 0% (0 of 43) in city 4 to 3.1% (1 of 32) in city 2.

The bacteria that caused the infections in the 12 patients included the following: 4 multidrug-sensitive Staphylococcus aureus infections (of which one was accompanied by Peptostreptococcus and another by Staphylococcus constellatus), 1 Haemophilus influenzae infection, 1 multiresistant Staphylococcus epidermidis infection, 1 Enterobacter cloacae infection, 1 Pseudomonas infection, and 1 Pseudomonas infection accompanied by Escherichia coli. No bacterium could be identified in 3 patients.

Five patients underwent implant removal. One explantation was the result of a malfunction after head trauma and occurred 5 months after the infection. The other 4 explantations were directly associated with infection. The first of those was explanted 2 months after the onset of infection (7 months after surgery), while the others were explanted immediately after the onset of infection.

Discussion

In this study of 1180 patients with cochlear implants across 8 different centers in France, we observed 12 postoperative major infections (1.0%). Postoperative infection was associated with perioperative antibiotic choice. Patients who received 48 hours or less of antibiotics (antibiotic prophylaxis) were almost 3 times as likely to develop infection as patients who received more than 48 hours of antibiotics (prolonged antibiotic treatment). Pediatric patients had almost 4 times the risk of developing infection as adults, and the infection rate among children who received antibiotic prophylaxis (3.2%) was almost 3 times higher than that observed among children who received prolonged antibiotic treatment (1.2%). These descriptive results suggest that prolonged antibiotic treatment in the pediatric population is reasonable in an attempt to reduce the risk of postoperative infection.

In the literature, several studies^16,17,18,19 have described complications that occurred after cochlear implantation, but few have investigated the role of antibiotic prophylaxis. It is also difficult to draw any definitive conclusions from small samples given the low infection rate. Therefore, the present series of 1180 patients, probably the highest number published to date in this field, is informative. Generally, the infection rate was low in the series, not exceeding 5%.

The first large series on this subject was published in 1989 by Robinson and Chopra.²⁰ A questionnaire was sent to 27 centers. A total of 1030 auditory implants were reported, but 300 devices were extracochlear implants, with 730 patients receiving cochlear implants. Antibiotics for 1 to 14 days were used in 56.4% of cases, whereas no antibiotics were used in the others. The rate of infections associated with explantation was 2.9% and was higher when antibiotics were used (4.9% vs 0.9%), leading the authors to conclude that using antibiotics was without benefit. That said, there was no group receiving a single prophylactic dose in their study, and there were more implants with extracochlear electrode arrays in the group receiving antibiotics.

In 2004, Basavaraj et al²¹ retrospectively compared a single dose against prolonged antibiotics in cochlear implantation among 292 adults and children receiving implants. During more than 15 years, 12 infections occurred (4.1%), 4 of which were major infections, as well as 2 explantations. The infection rate was equally high in both groups, the first one receiving prolonged antibiotics and the second one receiving a single prophylaxis dose, and was higher when there were comorbidities. In our series, we also found a comorbidity that likely promoted the infection in almost half (5 of 12) of the infected patients. However, because of the study design and sample size, it is impossible to recommend from our data a precise attitude regarding prescription of antibiotics in the presence of comorbidities. Among our cases of major infection with comorbidity, 2 patients received prolonged antibiotic treatment while 3 patients received antibiotic prophylaxis.

In 2007, Hirsh et al²² observed 3 minor infections (3%) and no major infections in a retrospective series of 95 patients (98 implants). Seventy-eight patients received a single perioperative dose of antibiotics, with no association with major infections. However, their small sample size prevented definitive conclusions.

Garcia-Valdecasas et al²³ retrospectively studied SSI in 196 patients between July 2005 and December 2007. They compared the following 2 protocols during 2 different periods: a prophylactic dose of ceftriaxone (96 patients before October 2006) and a prophylactic dose of ceftriaxone combined with clarithromycin for 6 weeks after surgery (100 patients after October 2006). They also compared implants with a ceramic casing vs implants with a titanium casing. In total, 9 major infections occurred (4.6%), 8 in the group that received a single dose of ceftriaxone and 1 in the group that received 6 weeks of ceftriaxone combined with clarithromycin. Therefore, the risk of infection was 8.1 times higher when a prophylactic dose of ceftriaxone was used compared with prolonged ceftriaxone combined with clarithromycin. All 9 infections resulted in explantation, and all occurred with titanium implants. The small sample size prevented definitive conclusions regarding their results.

Investigating a series of 294 implantations and 9 infections, a team led by McAllister and colleagues²⁴ has proposed reimplanting the same implant after disinfecting it and has reported good functional results. However, in our opinion, implanting back a nonorganic implantable medical device that has been bacterially contaminated, in close contact with the endolymph and consequently with the cerebrospinal fluid, is ill advised.

No cases of meningitis were observed in our study. This major life-threatening risk remains low. A resurgence in postimplantation meningitis was observed up to 2002, leading the US Food and Drug Administration to formulate several recommendations,^25,26 in particular vaccination against Pneumococcus and Haemophilus species.

Regarding microbiology, the bacteria identified herein are known to be opportunistic or associated with health care. The most prevalent bacterium was S aureus (5 cases). Precise recommendations apply to areas as varied as the perioperative decontamination of individuals carrying S aureus, cutaneous preparation of patients undergoing operation, disinfection of the hands, and even the maintenance of homeostasis (normal body temperature and glycemia).²⁷ Pseudomonas aeruginosa²⁸ and the enterobacters (E coli and E cloacae)²⁹ are also frequently identified in the literature as opportunistic. Staphylococcus constellatus³⁰ and Peptostreptococcus are rarer. Last, H influenzae, which was found in our study in a child aged 1½ years, is usually considered to be commensal, but it can be pathogenic and cause invasive infection.³¹

Most of the studies published to date on this subject had insufficient sample sizes or present heterogeneous series with different types of auditory implants or do not address the role of antibiotic prophylaxis,²⁰ making it impossible to draw any definitive conclusions. A recent literature review³² included only 3 of 167 articles initially identified. Most reports were excluded from the review because they did not focus on the role of antibiotics in the perioperative period or because they did not contain primary data but were only review articles or surveys. Published in November 2016, that review found insufficient evidence to draw any solid conclusions regarding the use of perioperative antibiotics.

Surgical site infection may occur years after cochlear implantation. Therefore, it is more difficult to calculate this risk than it is to calculate the risk of perioperative infection or infection within 1 year of implantation (as per the French National Authority for Health definition of infection attributable to an implanted device).³³

Limitations

Our study has some limitations. The main limitations are lack of randomization, masking, and standardization of interventions, as well as heterogeneity in the sample, duration of follow-up, and sample size. We do not believe that the care provided at the different centers was sufficiently heterogeneous to result in significant bias, with no significant association observed between the centers and our findings.

Conclusions

The infection rate after cochlear implant surgery is low. Fewer infections were seen in the group who received prolonged antibiotic treatment and in adults compared with children. Given the dearth of randomized clinical trials and reliable meta-analyses on this subject and in light of our large series of almost 1200 patients, we prefer to administer a single dose of antibiotic prophylaxis in adults and prolonged antibiotic treatment for 7 days in children.

Supplement.

Data Sharing Statement.