Introduction

Deep brain stimulation is widely recognized as a safe, effective and reversible treatment for a variety of movement disorders, such as Parkinson's disease, primary dystonia and essential tremor. Recently, DBS has been approved for obsessive-compulsive disorder and epilepsy, and new indications, such as for cluster headaches, are under investigation [1, 13].

Device-related infection is one of the most serious and difficult to manage postoperative complications of DBS implantation. Treatment of these complications may take a long time and may result in additional hospitalizations, surgery and partial or total removal of the DBS system. Hardware explantation may lead to a loss of efficacy of therapy and exacerbation of symptoms throughout the period of time that elapses before the system can be replaced. The reported incidence of infections after a DBS procedure varies widely, from 0.4 to 12.7% [1–13]. Prompt implementation of an effective management protocol is essential in order to deal successfully with device-related infections and to reduce the risk of total removal of the system.

The aims of this retrospective review of 212 DBS procedures were to evaluate the incidence of device-related infections, to investigate the clinical characteristics of infections and to describe a management protocol that could minimize the consequences of infections.

Method

Study design

Data on 106 patients, in whom 212 DBS procedures were performed between 2001 and 2011 at the Neurosurgery Department of San Bortolo Hospital, were reviewed to assess the incidence of infections in the first year after device implantation.

Demographic data (sex, age on implantation, diagnosis and comorbidities) and information about targets were collected for all patients included in the analysis, as was all information on infections. These data were obtained retrospectively, with the patients’ consent, from clinical records.

All procedures were performed by a single neurosurgeon (M.P.). Any infection that occurred after any DBS-related surgical procedure—whether for the implantation or replacement of leads or implantable pulse generator (IPG)—was included in the analysis. Only deep infections were considered in this review. Deep infections were defined as infections that extended into the subcutaneous layer and were usually in contact with at least one part of the DBS system. The infections were diagnosed on the basis of clinical or microbiological evidence. Clinical evidence included cellulites surrounding the hardware component or purulent drainage, while microbiological evidence consisted of positive bacterial cultures from the infected site.

The infection rate was calculated both as episodes/procedures and episodes/patients. In the first year after implantation, all patients were assessed every 3 months or at the onset of symptoms (pain, wound redness, swelling of the skin).

Surgical procedure

A preoperative antibiotic therapy consisting of antistaphylococcal cephalosporin (2–3 g intravenously) was administered to all patients prior to skin incision.

Between 1998 and 2006, preoperative planning was based on ventriculography and MRI images merged with CT images, while from the beginning of 2007 targeting was based on MRI images merged with angiographic CT by a computer-assisted neuronavigation system (Stealth Planning Station Medtronic, Inc., Minneapolis, MN).

All patients underwent two-stage surgery, in which leads were implanted during the first surgery session, and extensions and IPG were implanted a week later during the second surgery session.

A Cosman-Roberts-Wells frame (Radionics Inc., Burlington, MA) was used for the placement of all electrodes. The frame was placed on the patient’s head on the morning of the surgery. The leads were stereotactically inserted through bur holes, whose position on the scalp depended on the target’s trajectories. Lead insertion was performed under local anesthesia. The optimal electrode location was defined on the basis of planned coordinates, and the data obtained from micro-recordings and intraoperative stimulation. All body parts involved in all the surgical procedures were washed and disinfected with a betadine solution. In bilateral implants, the surgical field was removed, and a new one was made between the first and the second lead placement. Leads were secured by means of acrylic resin, transcranial suture and a titanium plate.

An IPG was implanted under general anesthesia during the second surgical procedure. The IPG was connected to each DBS electrode by means of an extension wire and placed in the infraclavicular region. The electrode connectors were placed in the mastoid area, and the frontal incision on the head was reopened during the second procedure in order to connect the IPG with the leads.

Oral antistaphylococcal cephalosporin (3 g) administration continued for 2 weeks after each surgical procedure.

Statistical analysis

Data are presented as averages. A Student’s t test was used to compare the ages of infected and uninfected patients. Indications, targets and comorbidites were compared between the two groups by using a χ2 test based on two-by-two contingency tables. All analyses were conducted by means of a two-sided test. A p <0.05 was considered significant. Statistical analyses were performed by means of SPSS 12.0 (SPSS Inc., Chicago, IL).

Results

A total of 106 patients (23 females and 83 males; mean age on implantation 55 ± 2.3 years) underwent DBS system implantation. In 102 patients, the implantation was bilateral, while 4 patients treated for cluster headaches underwent unilateral implantation. Ninety-six patients were affected by Parkinson’s disease, six by epilepsy and four by cluster headaches. Exfoliative dermatitis was the only comorbidity recorded: in four patients, all in the infected group.

In all parkinsonian patients, the STN was the target chosen; the anterior nucleus of the thalamus was chosen as the target in epilepsy patients and the posterior hypothalamus in patients with cluster headaches.

Postoperative infections occurred in 8.5% (9/106) of patients and in 4.2% (9/212) of procedures. The sex distribution and the mean age on implantation were not statistically different between the infected and uninfected groups (8 males and 1 female infected, p = 0.68; mean age on implantation: 55 years in the infected group and 54 in uninfected patients, p >0.05).

All patients suffering from exfoliative dermatitis incurred a postoperative infection, but there was no significant difference in the distribution of this comorbidity between the two groups (p > 0.05). All infections occurred in parkinsonian patients.

The mean time to the onset of infection was 30.7 days. In 7 patients, infection occurred within 30 days after surgery (early infection rate: 6.6% of patients and 3.3% of procedures); in 2 patients, infection arose within 6 months after implantation (late infection rate: 1.8% of patients and 1% of procedures).

All patients underwent two-stage surgery. The mean duration of the first stage of the procedure (lead placement) was 8 h, while IPG implantation lasted 52 min.

The mean interval of time between the two stages of surgery was 8.5 days (range 7–10 days).

Culture results were positive in seven infections (Table 1).

Table 1 Data and characteristics of all infected patients
Full size table

Seven infections involved the IPG site, the rate of IPG site infection being 6.6% of patients and 3.3% of procedures. In eight of the nine patients who suffered infections, the IPG had been placed in the infraclavicular region of the chest, while in the remaining patient it was in the left abdominal region. Two infections involved the connector site, the rate of infection at this site being 1.9% of patients and 0.9% of procedures. The rates of infection at the two sites were not significantly different (p >0.05). There was only one infection involving the bur hole, and no infection extended intracranially.

In eight cases of infection, the IPG and the extension were removed; they were re-implanted 3 months later, once freedom from infection had been ascertained. In the remaining patient, the infection extended to the bur hole, and the whole system, including the brain leads, had to be removed in order to avoid consequent intracranial infection. This patient refused to undergo re-implantation of the system.

An antibiotic therapy was administered to all infected patients for 4 weeks on the basis of culture and sensitivity results. If clinical results showed involvement of the hardware, patients underwent a surgical procedure for the partial or total removal of the system.

Discussion

Hardware-related infection is a common complication of DBS system implantation, with reported rates of infection varying between 0 and 15% [11]. In our retrospective analysis, infections occurred in 8.5% of patients and 4.2% of procedures. In a literature review carried out by Bhatia et al. [1], the cumulative overall rate of infections was 3.8% of patients and 3.3% of procedures. In their single-center experience, however, the same authors reported a higher infection rate, owing to the inclusion of both short-term and long-term infections, superficial and deep episodes, with or without hardware removal. The incidence revealed by the review of our DBS implantation procedures is in line with that reported by Bhatia et al. [1]. In our experience, infections involved IPG sites in 3.3% of procedures and the connector site in 0.9%. No infection occurred at the bur hole site, in agreement with the literature data [11]. The recent literature indicates that the infection rate is not influenced by patient age, sex, target or modality of IPG placement (a two-stage procedure vs. implantation simultaneously with leads [1, 11]; our data confirm this). Although Sillay et al. [11] showed that infection risk seems not to be influenced by the indication, in our cohort all infections occurred in parkinsonian patients. This may be due to the worse health conditions of these patients. In our infected group, the rate of exfoliative dermatitis was higher than in uninfected patients, though the difference was not significant. Factors that may influence the frequency of infections are the length of surgery, prophylactic administration of antibiotics, handling of the implants and the hygiene habits of the patients. In our patient cohort, three infections occurred in patients with poor personal hygiene and a low cultural background; however, the sample size was too small to establish a relationship between these aspects and the rate of infections.

In our patients, the duration of surgery seemed to be unrelated to the occurrence of infections. Indeed, all infections involved the stimulator and extensions, components that were implanted during the second, shorter stage of the surgery. Moreover, the interval between the two surgical stages of implantation could not have influenced the infection rate, as none of the infections occurred in wounds that were reopened and re-sutured during the second stage.

In our review, most infections (7/9) occurred in the first 30 postoperative days. We did not implement a standardized procedure for the management of device-related infections. Some reports have suggested complete removal of the whole system if the infection involves the bur hole, and partial removal of the system if the infection involves device components other than leads, in order to avoid new intracranial stereotactic surgery to replace the electrodes. Our current management strategy is to treat infections with local wound care and antibiotic therapy, without partial removal of the system. If clinical results show involvement of the hardware, or if the infection does not resolve within 4 weeks from the beginning of the treatment, the system is partially removed. Three months after removal of the system, patients are evaluated for re-implantation if the infection has completely resolved.

Conclusion

In a series of 106 patients who underwent 212 implantation procedures, the overall rate of hardware-related infections was 4.2%. Timely, effective management of the infections, which involved the IPG or connector site, played a key role in avoiding complete removal of the system.

Our review suggests that the only factor predictive of the occurrence of infection is the hygiene habits of the patients. Age, sex, target, IPG placement in the second stage (as opposed to simultaneously), duration of surgery and the interval between the two surgical stages seemed not to have any influence on the infection rate. These results suggest that the duration of antibiotic therapy should be reconsidered and shortened in order to avoid favoring the emergence of antibiotic-resistant virulent or saprophytic bacteria.

Our algorithm enabled us to limit the spread of infection and to partially preserve the system in 88.8% of cases. The preoperative administration of antibiotic therapy and the implementation of a therapeutic algorithm for the management of infections are essential for limiting the impact and severity of these complications.