Ultrasound-assisted treatment of sternocutaneous fistula in post-sternotomy cardiac surgery patients^†

Abstract

OBJECTIVES

Using an ultrasonic debridement system (low frequency, 25 kHz), we aimed to completely remove bacterial biofilms and preserve vital sternal tissue and to compare this procedure with conventional surgical therapy.

METHODS

In this retrospective study, we evaluated 37 consecutive patients (25 males) between April 2011 and June 2013 in whom sternocutaneous fistula (SCF) was treated with the ultrasound-assisted wound (UAW) system (Group A, n = 18) or with conventional surgical therapy (Group B, n = 19). Treatment in Group A consisted of a complete dissection of the SCF followed by a multistep UAW debridement session after an interval of 3 days. Our final step in both groups was secondary wound closure with a musculocutaneous flap.

RESULTS

Patients in both groups were categorized as high risk with respect to several of the known SCF risk factors. In both groups, a similar variety of bacteria were isolated: 61% were gram-positive species, 16.5% were gram-negative species and 10.5% were Candida albicans. Time to secondary wound closure following eradication was significantly shorter in Group A (10 ± 5.4 vs 15 ± 7.1 days in Group B, P = 0.012). Postoperative antibiotic treatment time (16 ± 9.3 vs 22 ± 10.7 days in Group B, P = 0.078) showed a trend in favour of Group A, but the mean hospitalization time (22 ± 12.0 vs 26 ± 14.3 days in Group B, P = 0.34) did not differ between groups. Recurrence of SCF tended to be less frequent in Group A (6 vs 21% in Group B, P = 0.46). In Group B, one infection-related death was noted. The mean follow-up time was 8 ± 2.7 (Group A) and 10 ± 5.7 (Group B) months.

CONCLUSIONS

Ultrasonic debridement is a promising adjunct to SCF treatment. In combination with adequate surgical and antimicrobial therapy, we documented good mid-term results in our trial group.

INTRODUCTION

In a patient population with growing numbers of multimorbid patients, post-sternotomy wound infections and mediastinitis are increasingly common. Despite several conservative and surgical reconstruction strategies, recurrent sternal wound infection (SWI, classified according to Pairolero and Arnold [1]) and sternocutaneous fistulas (SCFs) are still a serious problem. SCF is characterized by the presence of a sternal infection with slow onset; chronic draining of the sinus tracts and localized cellulites, osteomyelitis or retained foreign bodies. Mediastinitis is a rare finding in these patients, in whom infection is of low virulence without systemic reactions [1]. SCFs are rare but present a serious problem with high hospital costs and increased long-term morbidity and mortality independent of surgical mortality. The incidence rate of secondary SCF following previous deep SWI is between 3 and 10% [2, 3]. Some authors suggest that SCF has a multifactorial cause, including previous SWI, renal failure, smoking and use of excessive bone wax as major risk factors. In addition to the aetiology of SCF, there are further discrepancies between the conservative and surgical therapies for SCF. Some authors suggest that an exclusive antimicrobial therapy without surgical intervention represents the most effective treatment for SCF [2, 4–6]. Others suggest that surgical excision of the corpus alienum or bone sequesters, followed by vacuum wound sealing and a systemic antimicrobial therapy, is a safe and effective approach [1, 6, 7]. The clinical appearance of this suppurative infection differs from microbiological species to species: bacterial SCF differs from fungal infections. Some Staphylococci (epidermidis, coagulase-negative) and Candida species cause opportunistic infections that originate from foreign bodies. These microorganisms adhere to foreign material or necrotic sternal tissue, create a biofilm and lead to a low-virulence infection with slow onset [8–11]. Similar to both gram-positive and gram-negative bacteria, Candida species exhibit a propensity to grow as biofilms on implanted medical devices. Treating these infections proves challenging due to their high levels of drug resistance. Compared with their planktonic counterparts, biofilm-resident bacteria exhibit an up to 1000-fold increase in resistance [12–14]. Effective biofilm eradication without radical sternal bone destruction is not feasible in a conventional surgical approach. To optimize debridement and maximize the bactericide effect, ultrasound-assisted wound (UAW) debridement systems seem to be an effective alternative method. The treatment of chronic wounds with ultrasound achieves selective necrosectomy, a reduction of the bacterial burden and stimulation of granulation. These effects can be attributed to mechanical, mechano-acoustic, thermal and non-thermal stimuli [15–17]. This cumulative bio-acoustic effect permeates the bacterial cell membrane, and the absorption of antibiotics and antiseptics is increased in both gram-positive and gram-negative bacteria [18–21].

The use of low-frequency ultrasonic wound treatment has not yet been investigated in SWIs, and its relevance as an effective alternative method for SCF therapy remains unknown. Our hypothesis is that using an ultrasonic debridement system (low frequency, 25 kHz) will completely remove the bacterial biofilms, and better preserve vital sternal tissue compared with conventional surgical therapy alone. In addition to surgical advantages, we hypothesized that this technique will reduce hospital costs by shortening the length of the hospital stay and the length of antimicrobial therapy.

PATIENTS AND METHODS

Between April 2011 and June 2013, we performed 3840 median sternotomies for cardiac surgery procedures in our institution. Thirty-seven patients (0.97%) developed SCF. All patients who developed SCF after being discharged from the hospital after an uncomplicated primary cardiac procedure were eligible for inclusion in the study. Exclusion criteria were as follows: patients with a subcutaneous tissue fistula without sternal infection, patients with pericarditis near the aorta, patients with bacteraemia and multiorgan failure and patients with a post-percutaneous tracheotomy.

In this retrospective study, we compared all patients with SCF who were treated either with an UAW system (see detailed description below; Group A, N = 18) or with surgical debridement followed by vacuum assisted closure (VAC) therapy (Group B, N = 19). SCF was characterized by the presence of chronic draining sinus tracts localized in the sternotomy scar, with cellulites, sternum osteomyelitis, infected foreign bodies or sequestrations. Primary SCF was characterized by a diagnosis of a fistula months or years after a median sternotomy without previous SWI or mediastinitis. Secondary SCF included fistulas that developed weeks or months after a deep SWI or mediastinitis treatment. In both groups, SCF was diagnosed by clinical examination (local infection signs and fever) or laboratory findings (leucocytosis and C-reactive protein), and confirmed by a CT scan or fistulography. The diagnosis was confirmed by radiological examination in all cases. Microbiological wound cultures were collected prior to wound debridement. Two sets of wound swabs were sent for immediate gram stain and culture. In cases with negative wound culture, wound tissues (soft or bone) were collected for PCR assay. Once an SCF became evident, appropriate antibiotics were administered based on culture and sensitivity results. Sternum osteomyelitis was diagnosed when bone morphological changes were documented by a CT scan, or were noted at the time of sternal resection or removal of wires or sequester. Prior to SCF debridement, informed consent was obtained. The treatment decision was surgeon-based, and patients with SCF protruding into the pericardium were excluded from the ultrasound treatment group as the effect of the UAW device on the myocardium is unknown. General anaesthesia was practised in all patients.

Statistical analysis

All values are presented as the percent of treatment group, with absolute numbers in parentheses, or as the mean plus/minus standard deviation. The statistical analyses were performed in SPSS 22 (SPSS, IBM Corporation, Armonk, NY, USA). Student's t-test was performed where normal distribution of data was estimated. Otherwise, a Mann–Whitney test was performed. Survival analysis was performed for overall mortality and recurrence of SCF. A value of P < 0.05 was considered statistically significant.

RESULTS

Primary SCF as a complication of post-median sternotomy was found in 67% (12/18, Group A) and 79% (15/19, Group B) of the patients. In Group A, 33% (6/18) of the patients developed SCF after SWI and mediastinitis; in Group B, 21% (4/19) similarly developed SCF. The patients' characteristics and perioperative risk factors are summarized in Table 1. The mean ages were 65 (±12) (Group A) and 68 (±9) (Group B) years. Group B consisted of 16% female patients compared with 50% in Group A. In both groups, the patients presented a multitude of known SWI risk factors (e.g. nicotine abuse, bilateral internal mammary artery (IMA) harvesting and use of bone wax). In 56% of patients in Group A, bilateral internal thoracic artery harvesting was performed. The mean operation time exceeded standard times in both groups (Group A: 216 ± 62 vs Group B: 209 ± 70 min). In both groups, bone wax was used excessively. Between groups, only gender (P = 0.038), bilateral use of IMA (P = 0.0001) and cardiopulmonary bypass time (0.046) differed significantly. Other risk factors (diabetes, BMI, COPD, chronic kidney failure, peripheral arterial disease, CABG, operation time, use of bone wax and log EuroSCORE) were similar in both groups. There was no difference in the cardiac surgery procedures that were performed in both groups. The stay in the ICU was twice as long in Group B compared with Group A (P = 0.021). The mean time until the SCF diagnosis was similar in both groups, approximately the 25th week post-median sternotomy.

The UAW (Sonoca 185, Söring, Inc., Quickborn, Germany) system consists of a generator and an ultrasonic transducer (Fig. 1). Normal saline (0.9% NaCl) was used as the connecting medium for the ultrasound waves to the tissues, and was procured by a transducer-attached line. The transducer and a foot pedal were connected to the generator. The low frequency (25 kHz) allows for great depth of penetration, with only minimal thermal stress to the tissue. In contrast, the high intensity (35–40 W/cm²) creates a strong cavitation effect that is essential for the therapeutic effect. Cavitation is defined as the formation of microbubbles in a liquid medium induced by ultrasound-induced compression/traction forces. This technique allows the wound to be cleaned of necrotic tissue while suppressing the development of surface microflora. At this power level, exposure to ultrasound is not destructive to healthy tissues, and selectively removes only pathologically altered tissue. Without causing extra tissue damage, necrotic wound tissue is separated from the vital tissue. With a force setting between 60 and 100% of the device's maximal power, the wound surface is treated for 15–30 s per cm², as recommended by the manufacturer. All foreign materials in and around the SCF area, infected scar and bone sequesters were removed with standard surgical instruments. UAW was used to dissect the SCF; this step was followed by cleansing the surrounding vital tissue and performing wound compression with sterile gauze. Wound care on a daily basis consisted of wound irrigation with normal saline (0.9%) and wound compression with sterile gauze. This UAW debridement consisted of multiple sessions with a standard interval of 3 days (an example of primary SCF is presented in Figs 2 and 3). After complete eradication, the sternal defect was closed with a musculocutaneous flap.

Conventional surgical debridement consisted of excision of the fistulous tracts together with their blind ends, removal of foreign material or sequestration followed by temporary vacuum-assisted closure or the placement of absorbent cotton gauze with iodoform (Lohmann & Rauscher International GmbH & Co., Germany). VAC therapy was continued until the wound appeared macroscopically clean, granulation tissue had formed and the microbiology results revealed no remaining bacterial contamination. Our final step in both groups was secondary wound closure with a musculocutaneous flap.

In approximately 90% of the bacteriological assays in both groups, a variety of bacteria were isolated. In Group A, 61% (11/18) were gram-positive, 17% (3/18) were gram-negative bacteria and 11% (2/18) were Candida albicans. In Group B, 63% (12/19) were gram-positive, 16% (3/19) were gram-negative bacteria and 11% (2/19) were C. albicans. There was no significant difference between groups in terms of the isolated bacterial and fungal species (Table 2). A difference was observed between the primary and secondary SCF patient populations (Fig. 4). In the primary SCF, the bacteriological assays consisted mostly of multiresistant bacterial or fungal species. These microorganisms can create a biofilm and survive longer, with a slow onset.

After the second UAW debridement session, we found 89% (16/18) clearance of bacteria in wound cultures. In 2 (11%) patients, a third session was needed. After the third session, we detected C. albicans DNA using the PCR assay in 1 patient's wound tissue. In this patient, a radical surgical eradication was needed due to therapy-resistant C. albicans.

The mean UAW debridement session lasted 14 (±4) min. The postoperative data are summarized in Table 3. VAC therapy was needed in 15 patients in Group B, with a mean VAC therapy time of 10 (±7) days. In Group A, the sternal wound was sealed with sterile gauze after every UAW session. Compared with Group B, secondary wound closure was achieved earlier in Group A [A: 10 (±5) vs B: 15 (±7) days; P = 0.012]. Although not reaching statistical significance, the initial postoperative antibiotic therapy showed a trend towards longer treatment durations in Group B compared with Group A (P = 0.078). The mean hospitalization time did not differ between groups. The recurrence of SCF was higher in Group B (21 vs 6% in Group A; P = 0.46) without reaching statistical significance. One infection-related death was noted in Group B; after complete sternal wound healing, the patient was readmitted to the ICU 6 months post-wound therapy. He developed a septic cerebrovascular infarction with right hemispheric stroke. In blood samples, C. albicans DNA was isolated. On a transoesophageal echocardiogram, an aortic valve prosthesis endocarditis was found. The patient died a few days later in septic multiorgan failure. The mean follow-up time was 8 (±3) in Group A and 10 (±6) months in Group B.

DISCUSSION

To date, this is the first study to describe the effect of ultrasonic sternal wound lavage in cardiothoracic surgery. The therapeutic improvement in comparison with conventional surgical debridement stems from the direct wound-cleansing effect. To determine the indication for one treatment compared with the other, we must clarify the aetiology and prevalence of SCF in post-sternotomy cardiac surgery patients. More than 80% of our trial patients presented after a significant delay from their initial surgery, and were referred from a wide variety of specialists and often after several errant regimens of narcotics, anxiolytics or antidepressants that failed to ameliorate their symptoms. In cardiac surgery, patient profiles have changed dramatically in the last two decades, and more than 70% of patients must currently be categorized as being at high risk for the development of SWIs and SCF. Most patients present with a multitude of risk factors in addition to iatrogenic risk factors such as excessive use of bone wax and electrocauterization of the sternal periosteum. Both types of risk factors are responsible for the development of impaired sternal bone healing, which results in sternal non-union and pseudoarthrosis [3, 4]. The latter in particular can develop without any external signs or symptoms of sternal infection. In these patients, the risk of sternal infection is highly underestimated. Staphylococci species (epidermidis and coagulase-negative) and C. albicans can survive by creating a biofilm, and adhere to foreign material or devitalized sternal bone. In this low-virulence, therapeutic-resistant state, the microorganism can survive for several months or years. A diagnosis of chronic sternal infection in the first 6 months post-sternotomy is difficult. In this period, differentiation of pseudoarthrosis and sternum fistula is also difficult. In a prospective CT scan study of sternal healing after median sternotomy, Bitkover et al. [22] concluded that there was no sternal healing 3 months after the operation, and even 6 months postoperatively, complete healing was only achieved in half of the patients. With regard to draining sinus tracts that are surrounded by localized cellulites, a CT fistulogram provides detailed information, and is often superior to simple X-ray fistulography. This approach can be particularly helpful in confirming the diagnosis and identifying the number and location of SCFs, sequesters, steel wires or retrosternal abscess and planning the surgery. SCF debridement with conventional surgical techniques (Group B) presents several disadvantages: (i) the use of sharp instruments for debridement can either affect deep vital tissues or (ii) will not sufficiently decrease bacterial flora. (iii) Using negative pressure therapy (e.g. VAC systems) is easy but time-consuming. (iv) The removal of necrotic tissue alone will not completely reduce the bacterial biofilms. (v) Iatrogenic bacteraemia with the risk of acute septic shock is very common in such procedures.

In the conventional surgical treatment group, we found a higher number of SCF recurrence. Ultrasound as a therapeutic agent in chronic wound healing has been extensively studied. The overall acceptance in other medical disciplines (e.g. septic, vascular and orthopaedic surgery) is very good, especially for soft tissue and bone debridement. The healing process is much faster in acute infections and in chronic wounds [20, 21, 23, 24]. Based on the cavitation effect, the bactericidal properties of ultrasonic lavage offer an effective treatment for infections, and sufficiently eliminate biofilms [17, 23]. We accordingly observed a faster secondary wound closure time following ultrasound treatment in our study.

The tissue-selective properties of ultrasonic debridement allow for a very gentle treatment. Only necrotic tissues, cell remnants, biofilms and contaminations are safely removed, and vital tissue is hardly affected. However, not all necrotic bone tissue or infected steel wires can be removed with UAW alone; a selective surgical debridement is still needed. A combination of conventional surgical debridement techniques and ultrasonic lavage is most effective. The main goal of our treatment strategy was to improve the healing process by maximally reducing the bacterial tissue load. In complex sternum bone fistulas, we needed surgical instruments to achieve maximal lavage in the deepest sinus ends. However, no surgical preparation was able to destroy biofilms as effectively as ultrasonic lavage. Despite the cited advantages of ultrasonic lavage, we also documented a few minor disadvantages of the UAW system during tissue contact. (i) The tissue temperature rose slightly during the treatment with the UAW system. Tissue overheating can lead to necrosis. In our patient group, sternal tissue overheating was prevented by continuous saline irrigation. (ii) The possibility of contamination or transmission of disease by infectious saline mist produced by the transducer in patients with resistant viral infections (e.g. HIV, hepatitis and methicillin-resistant staphylococcus aureus) is unknown. (iii) The UAW system transducer (sonotrode) contact to steel wires might lead to a complete transducer defect. (iv) Not all necrotic tissue can be removed by utilizing the UAW system. For infected scars or devitalized bone, a surgical instrument is still needed. (v) For the ultrasonic SCF treatment, general anaesthesia is mandatory. (vi) The high treatment costs and the need for transducer sterilization after every session are additional limitations. However, ultrasonic therapy is easy to learn and time-saving, and shows a favourable risk profile. As chronically implanted materials (e.g. wires, plates and clamps) become more common during cardiac surgery, the concern with respect to the contamination of these medical devices grows. Although the incidence of such infections may be low, and continues to decrease, the results are often severe and even fatal to the patient. The bacteria that grow in biofilms on synthetic medical implants are highly resistant to traditional antimicrobial therapy. This recalcitrance necessitates the surgical removal of infected implants, followed by ultrasonic lavage and systemic antimicrobial therapy to reduce bacterial growth and successfully treat the associated infection. The effect of these three clinical treatment regimens will facilitate early wound healing, which will result in a reduction of total healthcare costs, and improve the patient's quality of life.

Study limitations

The present study suffers from the shortcomings of a small, monocentric retrospective study. The fortunately low incidence of SCF only allows for small patient numbers per centre. A multicentric, prospective approach is the next step forward following the promising results of this pilot study.

CONCLUSIONS

SCFs are caused by bacterial or fungal species that can develop biofilms and resistance against most conservative treatment strategies. The disease can prolong hospitalization, and increase medical costs. The UAW system is an effective alternative tool for the treatment of bacterial SCF; however, it seems to be less effective for the treatment of C. albicans-infected fistulae. In our opinion, ultrasonic debridement is a promising adjunct to surgical SCF treatment. In combination with adequate surgical and antimicrobial therapy, we documented good mid-term results in our trial group.

Funding

This study was supported by a research grant from Söring, Quickborn, Germany. The sponsor was not involved in any phase of the study preparation, data collection, data analysis or manuscript preparation.

Conflict of interest: none declared.

REFERENCES

APPENDIX. CONFERENCE DISCUSSION

Scan to your mobile or go to http://www.oxfordjournals.org/page/6153/1 to search for the presentation on the EACTS library

Dr H. Pilegaard(Aarhus, Denmark): I think your manuscript raises several questions. When you look at sternal infections which is a very important field and topic, because it costs a lot of money if the patient gets an infection after cardiac surgery or another kind of surgery where you need a sternotomy. You have two groups, and I think the only thing which is different in the two groups that you have, is ultrasound in one group; and use of VAC in the other group, because I think you need the surgical intervention in both groups.

But even though you tell us that there is no difference, I think the statistics are a little bit weak, because they are small groups. I would like to know how you have decided the treatment of the patients. Because when we look at your groups, in one group, you have all bilateral mammary arteries, the shortest pump time; and in the other group, you have the longest pump time and no bilateral internal mammary use. How did you decide to treat the patients? Was it the surgeon's choice or what did you look for?

Dr Tewarie: I think it was more the surgeon's choice, because we do not know what the effect of ultrasound will be on myocardial tissue and on bypass graft.

As I said, this device has been used in neurosurgery, and in septic orthopedic surgery. They have used it plenty of times, but they did not mention anything about vascular damage or brain tissue damage. I was very careful, as I did not know what the effect will be, so I choose per patient. When there was a secondary sternocutaneous fistula preceding a deep sternal wound infection or mediastinitis, I did not use this device, because I have to go in the pericardium, and make the debridement near the myocardial tissue. So, I did not use this system in those kinds of patients, and I think that is the reason why there is a difficulty in both groups.

Dr Pilegaard: But this number might be very small because you tell us that there are only patients with fistulas. There is no mediastinitis. So why did you have to go into the pericardium?

Dr Tewarie: The secondary fistula was a patient with mediastinitis and deep sternal wound infection. So, we do not know if we also need a CT scan. A fistulogram is not sensitive in the first few months, and a CT scan is not sensitive for the first six months until one year, there is no sensitivity of CT scan, so it is difficult.

If you make that choice and you have to go to the pericardium or myocardial tissue, I do not know what the effect will be. So it can be dangerous.

Dr Pilegaard: Okay. Another thing, if you compare the two groups, in Group B where you use a VAC system, you have a much longer ICU time, and it is very expensive to have the patients in the ICU. But had you decided before you started the treatment that only patients with the VAC should go there, then I do not think it is fair to compare these two groups and to say one group can go direct to the normal ward, and the other group has to go to the ICU. Of course you will have a significant difference between the two groups.

Dr Tewarie: You are right, because the decision was made in the operation room.

Sometimes we have patients in whom we have to open the sternum and we are draining the pericardium, those patients also have a huge bleeding problem. They are not only patients of post-CABG, but they are patients who have had implantation or mechanical valve implantation. These patients have to be monitored in the ICU in our hospital. So, we send the patient for one day, or sometimes two days in the ICU, and afterwards, they go to a normal ward.

Dr Pilegaard: Another thing I wonder is when you look at the two groups, when you have the ultrasound group; you save six days because you can close the wound six days before the other group, median.

Dr Tewarie: 10 days. The mean time was 9.

Dr Pilegaard: 9 days in the ultrasound and 15 days in the VAC group. But then when you look at their stay in the hospital, there is no significant difference why?

Dr Tewarie: The problem is the wound, and for the patient treated with the ultrasound device, it is very small.

So, after the eradication in both groups, when you look to the micro-bilateral cultures, and when there is a complete eradication of the bacterial or fungal species, we close the wounds. But in the conventional surgical group, we would need more time, because during this procedure, you have the risk of the wound becoming contaminated. So it was not easy to close after two or three sessions in the conventional group.

Dr Pilegaard: Okay. You write in the paper that the treatment with the ultrasound is quite expensive.

Dr Tewarie: Yes.

Dr Pilegaard: Can you explain that?

Dr Tewarie: Yes. Only the generator will cost around 65,000 Euros, and the transducer I think is 10,000 Euros. The problem with this device is, as I said, it is only used in septic surgery and neurosurgery.

In our patients, we have also sternal vials, and when you have contact with this transducer to a steel vial, you will get a defective transducer, and it will cost a lot of money. So you have to take many precautions to prevent transducer defect.

Dr Pilegaard: Is it then cost effective to use ultrasound treatment?

Dr Tewarie: Yes.

Dr Pilegaard: That is cost effective?

Dr Tewarie: The generator is not a problem, the problem is the transducer.

Dr Pilegaard: Okay. I think, you need further studies in this field, and you need to have a randomized study comparing the two groups.

Author notes