Closure of Median Sternotomy in Dogs: Suture Versus Wire
Twenty normal, large-breed dogs underwent median sternotomy. Median sternotomies were closed with 20-gauge orthopedic wire in 10 dogs and no. 2 polybutester in 10 dogs. Closure with suture was faster than with wire (6.7±1.8 minutes versus 9.1±1.9 minutes, respectively). Significant differences were not observed in degree of postoperative pain or wound complication rates. Sternotomies closed with wire showed a trend to be more stable and had significantly less displacement on radiographic evaluation at 28 days. All sterna closed with wire examined histopathologically showed evidence of chondral or osteochondral bridging, while sterna closed with suture only showed fibrous union.
Introduction
Median sternotomy is a useful surgical technique in the dog, as it allows access to all thoracic structures.1–3 Closure of the sternotomy incision is achieved by stabilizing the sternal halves and reapposing overlying tissue. Traditionally, sternal closure in dogs has been accomplished using wire.1–5
Median sternotomy in the dog is associated with a high incidence of postoperative complications, ranging from 17% to 78%.6–9 These include postoperative pain, wound healing problems (e.g., incisional edema, discharge, seroma formation, and dehiscence), and osteomyelitis.6–9 This contrasts to complication rates of 0.5% to 5% associated with median sternotomy in humans.610–12 Anatomically, humans have a broader sternum, making a true midline sternotomy easier to perform. This has been identified as a necessary element to decrease postoperative complications.10–13
Recent reports in the human literature have described the use of various suture materials, including monofilament, multifilament, ribbon, and tape suture for sternotomy closure.14–19 The use of these materials is associated with an easier closure technique and a similar postoperative complication rate when compared to the use of surgical wire.
The use of suture material to close median sternotomy in the dog and cat has been reported in the veterinary literature.7–9 Polypropylene suture was used for the sternal closures in these reports. Williams, et al. had a 17% postoperative complication rate in a series of 18 closures;9 dog weights were not reported. Either wire or polypropylene was used in the report by Ringwald, et al.;7 however, complications associated with the different materials used for sternal closure were not reported.
Following sternotomy in humans, patients are often instructed not to raise their arms to shoulder height to avoid increased motion at the sternotomy site.2021 However, dogs must use their forelimbs following sternotomy for ambulation. Therefore, it is logical to assume that dogs place more stress on the sternum than humans do, and it would be prudent to evaluate alternative closure material on dogs before recommending its use.
This study directly compared two materials for median sternotomy closure. The objective was to determine if there was a difference in the complication rate and degree of postoperative pain between dogs in which the median sternotomies were closed with suture versus wire.
Materials and Methods
The research protocol was reviewed and approved by the Colorado State University Animal Care and Use Committee. Twenty dogs (age, 2 to 4 years; mean weight, 29±3 kg) involved in concurrent research requiring median sternotomy served as the study population. The dogs were randomized into one of two treatment groups for median sternotomy closure: 10 dogs underwent suture closure (no. 2 polybutestera), and 10 dogs underwent wire closure (20-gauge orthopedic wire). All dogs had a preanesthetic complete blood count, serum biochemical profile, and heartworm serology performed. Additionally, baseline heart and respiratory rates and pain threshold measurements from the sternum were obtained daily for 1 week prior to surgery. Pain threshold was measured using a load cell.b Increasing pressure was applied on the cell in a dorsal direction at the mid sternum to obtain a maximum tolerable force in Newtons (N). The measurement was recorded when the patient was perceived to react negatively to the stimulus or when a maximum force of 50 N was obtained. Negative reactions included vocalizing, attempting to move away, and turning the head toward the load cell.
All dogs were anesthetized with a standardized protocol. Acepromazine (0.02 mg/kg body weight) and morphine sulfate (1 mg/kg body weight) were administered subcutaneously as premedicants. Anesthesia was induced using intravenous (IV) diazepam (0.1 mg/kg body weight) and propofol (4 mg/kg body weight to effect). A surgical plane of anesthesia was maintained using an IV fentanyl constant-rate infusion (CRI) of 20 μg/kg body weight per hour and isoflurane in oxygen. Intermittent positive-pressure ventilation was used prior to entering the pleural space. A Swan-Ganz catheter and an arterial catheter were placed to monitor hemodynamic parameters intra- and postoperatively. These included systolic, diastolic, and mean arterial pressures; mean pulmonary arterial pressure; central venous pressure; cardiac output; and arterial and mixed-venous sampling for blood gas measurement.
Median sternotomy was performed using a sagittal sawc from the second sternebra through the xiphoid. The manubrium was left intact in all dogs. As part of the concurrent research, dogs underwent coronary arterial bypass grafting. The left internal thoracic artery was anastomosed to the left coronary artery in an end-to-side fashion. Prior to closure, a tube thoracostomy was placed through an intercostal space in the right chest wall. One surgeon (Monnet) performed all median sternotomy closures using an overlapping figure-of-eight pattern, placing a total of six wires or sutures per sternum, with an additional simple interrupted suture around the xiphoid if needed. Overlying muscle was closed with 2-0 polyglyconate in a simple continuous pattern. Subcutaneous tissues were apposed using 3-0 polyglyconate in a continuous horizontal pattern. Skin was closed with 3-0 nylon in a simple continuous pattern. The durations of median sternotomy closure and surgery were recorded. All dogs initially received postoperative analgesia at extubation, consisting of 5 mg fentanyl transdermal patch placement, a 2 μg/kg body weight IV fentanyl bolus, followed by immediate institution of a 4 μg/kg body weight per hour IV CRI of fentanyl. The fentanyl CRI was increased as needed, up to 10 μg/kg body weight per hour. Increases were based on clinical assessment of perceived painful behavior. Such behavior included vocalizing; unwillingness to sit, lie down, or rise; tachypnea; and agitation.
Dogs were assessed for pain at 1, 2, 4, 12, and 24 hours postoperatively using a combination of subjective and objective measurements [see Appendix]. A blinded observer (Halling) made all pain assessments. A multiple-category numerical rating system was used.2223 Categories observed were comfort, appearance, movement, unprovoked behavior, interactive behavior, heart rate, and respiratory rate. Pain threshold measurements were recorded. The amount of fentanyl required during each time period was recorded. Hemodynamic parameters were recorded at 4 and 24 hours postoperatively.
Incisions were examined daily, and wound healing complications were recorded. Pain threshold measurements were taken weekly. The endpoint of the study was 7 days for six dogs (three suture, three wire) and 28 days for 14 dogs (seven suture, seven wire). These dogs were humanely euthanized with IV sodium pentobarbital (90 mg/kg body weight). The sternum was evaluated grossly and deemed to be either stable or unstable based on apparent movement on manipulation. A sternum was deemed unstable if >2 cm of displacement could manually be obtained between the halves in a dorsoventral direction. Following euthanasia in the 28-day treatment group, the sternum was harvested and split transversely between the fifth and sixth sternebrae. The cranial portion was frozen at −70°C for radiographic evaluation, and the caudal portion was fixed in 10% neutral-buffered formalin for histopathological evaluation. Radiographs were taken in lateral and ventrodorsal positions. Displacement was measured in millimeters in each dimension (i.e., craniocaudal, sagittal, dorsoventral) and recorded for the third and fourth sternebrae. Degree of lysis and amount of healing were scored using a numerical rating system.
The unpaired Student’s t-test was performed on the following variables: age, weight, duration of median sternotomy closure, duration of surgery, average amount of fentanyl needed, and radiographic displacement. A two-way analysis of variance (ANOVA) for repeated measures was used to determine the effects of treatment, time, and treatment-time interactions for all other pain variables and hemodynamic parameters. A posthoc power analysis was performed on all pain and hemodynamic variables. The Fisher’s exact test was used to determine differences in wound healing complication rates, sternal stability, and sternal fixation breakage. The Mann-Whitney U test was performed on the degree of lysis and amount of healing based on radiographic evaluation. Variables were assigned significance when P<0.05. Statistical analyses were performed using a computer software package.d
Results
Surgery
There were no differences among treatment groups regarding age or weight. All dogs survived the surgery and study period. Median sternotomy closure was performed uneventfully in all cases. All median sternotomy closures were stable at the time of surgery based on inability to cause movement between the sternal halves on manipulation. Suture closure of the median sternotomy was significantly faster than wire closure (6.7±1.8 minutes versus 9.1±1.9 minutes, P=0.01). There was no difference in duration of surgery.
Postoperative Pain
None of the variables measured were found to be significant with respect to treatment [Table 1]. Seven variables were found to be significant with respect to time. Degree of movement and respiratory rate both increased with time. Unprovoked behavior score, heart rate, and arterial partial pressure of carbon dioxide (PaCO2) all decreased with time. The amount of fentanyl required as a CRI initially increased with time, then decreased. Additionally, there was not any significant interaction among these variables, with power calculations being >0.8 for all variables except for PaCO2 (power, 0.7).
All dogs followed a similar clinical course. All dogs remained in the critical care unit for 2 days postoperatively for thoracostomy tube and pain management. Dogs then returned to their hospital runs. All dogs were clinically normal based on behavior, food and water intake, degree of mobility, and lack of resistance on surgical site manipulation.
Wound Healing Complications
Overall, eight (40%) dogs developed wound healing complications. Complications occurred in both groups. Three dogs (one suture, two wire) had skin dehiscence. Underlying subcutaneous closure was not affected. One of these dogs (suture) also had prolonged incisional drainage for 5 days (day 3 to day 7, postoperatively). The draining fluid was serosanguineous and cytopathologically composed primarily of erythrocytes and nondegenerate neutrophils. Aerobic and anaerobic culture failed to demonstrate bacterial growth. This dog was empirically placed on enrofloxacin (5 mg/kg body weight, per os [PO] every 12 hours) and amoxicillin (22 mg/kg body weight, PO every 8 hours) pending culture results. Antibiotics were continued for a total of 7 days, 2 days beyond resolution of the incisional drainage. Five dogs (three suture, two wire) developed seromas ranging from 4 cm to 10 cm in diameter. All dogs were successfully treated by conservative management. No differences in overall or specific wound healing complication rates were found based on the Fisher’s exact test (P=0.99).
Necropsy Findings
All dogs had stable sternal fixations at 7 days. Two dogs with suture closure each had one broken suture. In the dogs with wire median sternotomy closure, no broken wires were observed.
Of the dogs in the 28-day group, all of the wire median sternotomy closures were deemed stable. One dog had one broken wire, and one dog had two broken wires. Three of the seven suture median sternotomy closures were deemed unstable. Two of these dogs had complete failure of all sutures, and the other had failure of four of six sutures. One to two sutures were broken in each of the remaining four suture median sternotomy closures. The difference in sternal stability was not significant (P=0.19).
Sternal fixation breakage was further assessed by visual inspection of the cranial portion (four figure-of-eight fixations) for the dogs in the 28-day treatment groups, for a total evaluation of 28 fixations in both groups. Two wires and 17 sutures were broken. The difference was significant as determined by the Fisher’s exact test (P<0.001).
Radiographic Findings
Five sternal halves from each treatment group in the 28-day duration group were radiographed [Figure 1]. No sterna were radiographed from the 7-day duration group. Radiographs of the sternal segments confirmed implant failures found at necropsy. Wire fixation resulted in a significantly smaller fracture gap in all planes . Indices were also assigned for amount of lysis and degree of osseous healing, with no significant difference found between treatment groups (P=0.29 and P=0.43, respectively).
None of the sterna showed complete osseous healing, although bony proliferation was evident on all sternebrae. Lysis was observed in two sternebrae in the same dog in the wire treatment group, with >50% of the sternal halves having undergone lysis.
Histopathological Findings
Three samples from each group were available for histopathological evaluation [Figure 2]. Suture closure resulted in more fibrosis and inflammation than did wire closure. Chondral or osteochondral bridging was not observed. This was in contrast to the wire closure group, where one specimen showed bridging of the sternotomy by osteochondral healing and the other two showed chondral bridging.
Discussion
This study supports the conclusion that wire closure results in better healing of the median sternotomy than suture closure. Implant failure was significantly greater in the suture closure group. This led to greater radiographic displacement and an increased incidence of instability at 28 days. Additionally, the limited number of samples from the suture closure group examined histopathologically showed marked inflammation within fibrous tissue. No chondral or osteochondral healing was observed. This is thought to be due to chronic instability. In contrast, the wire closure group had minimal inflammation and showed evidence of osteochondral bridging in one sternum and chondral bridging in two sterna. This supports wire closure as being more stable, resulting in less strain at the sternotomy site and subsequently resulting in better healing for dogs of this weight.
Both groups had wound complications of similar types and incidence as previously reported.6–9 No difference was noted between groups. Additionally, there was no clinically appreciable difference in outcome. All dogs showed similar behavior during the postoperative period.
There was no difference in postoperative pain assessments between treatment groups. Measuring pain in animals is problematic.2223 A multiple-category simple numeric scale was chosen for subjective evaluation of the dogs. Multiple categories provide a greater range of assessment over more simple measurements, such as single-category simple numerical scale or the visual analog pain scale.2324 Evoked pain response has been used previously as an objective measurement of pain.25 Objective measures of pain were not found to be different between treatment groups. Evoked pain response and hemodynamic parameters were similar between groups. Conzemius, et al.24 found a weak correlation between mean arterial pressure and subjective measures of pain. The results of this study support those findings.
Analgesic administration was based on individual pain assessments as opposed to using a standard protocol for both groups. This method was chosen for humane reasons and to reproduce a clinical situation. Differences in degree of postoperative pain should still be able to be easily recognized based on evoked pain response in comparison with the amount of analgesic required. Material used for sternotomy closure does not appear to have an influence on degree of postoperative pain, as evoked pain responses and fentanyl requirements were not significantly different at any time point in this study.
Pain assessments were not evaluated beyond 24 hours. They were not continued because of the variable duration of tube thoracostomy placement. The thoracostomy tube was maintained in all dogs for a minimum of 24 hours but were left in some dogs for up to 72 hours. As thoracostomy tubes are known to be painful, further pain assessment analysis would tend to be biased based on the presence of the thoracostomy tube and not a true reflection of the influence of sternotomy closure.
The suture material selected for use in this study was no. 2 polybutester. This selection was based on its strength, ease of handling, and knot security. It is a monofilament, nonabsorbable suture commonly available in private practice.
The overlapping figure-of-eight pattern was chosen as the technique for sternotomy closure. This technique avoids direct perpendicular shearing forces of the closure material and also aids in reinforcing the abaxial sternal segments.26 These two factors have been identified as critical to successful closure in humans.
This study has three limitations. The first is the duration of the study. Twenty-eight days is not long enough to completely assess sternal healing. As these dogs were involved in a concurrent project, a longer duration study was not possible. However, 28 days is a long enough period to observe for differences in postoperative pain and recovery rates.
The second limitation is the small sample number. A larger sample size may have been able to further identify differences in the treatment groups, especially at 28 days. This is especially true for assessment of stability and bone healing. However, at 28 days, the number of broken sutures was significantly higher than the number of wires broken. If the sample size was increased, it would be more likely to expect a more significant difference. Posthoc power calculations were performed for all pain and hemodynamic variables. With the exception of PaCO2, all power calculations were >0.8. Statistical power is the probability of having made a correct decision when the statistical tests reveal insignificance (P>0.05) and the null hypothesis is not rejected. The higher the power, the greater the chance that the decision is correct and no difference actually exists. Power values >0.9 support having made a correct decision in not rejecting the null hypothesis.27 Power values are a function of sample size and standard deviation. Since most power values for pain and hemodynamic parameters were >0.9, the authors concluded that the sample size was sufficiently large.
The third limitation is the ability to perceive pain in animals. It is interesting that so many dogs in the 28-day suture group had grossly unstable sterna, yet were not perceived to be in pain. This could be a reflection of the author’s inability to accurately quantify pain or differences in behavior displayed by these unsocialized research hounds.
Suture closure was significantly faster than wire closure, but only by a mean of 2.4 minutes. An experienced surgeon performed all sternotomy closures. The time difference may be greater with a less experienced surgeon, as wire closure is somewhat more awkward. However, once mastered, this technique results in only a short increase in time of closure that is thought to be inconsequential, considering the length of an average thoracic surgery.
Conclusion
This study directly compared closure of median sternotomy in the dog with suture versus wire. Wire closure was found to be superior and recommended for dogs of this weight (average, 29±3 kg body weight). Although a clinical difference was not observed, wire closure is preferred as it provides greater stability for greater duration, resulting in faster healing of the sternebrae. A longer duration study is warranted to further investigate these results.
Novafil; Davis and Geck, St. Louis, MO
Spring Action Load Device; Pain Diagnostics and Thermography, Great Neck, NY
Maxidriver; 3M Corporation, Minneapolis, MN
StatView 4.0; SAS Institute, Cary, NC
Acknowledgments
The authors acknowledge the assistance of Ms. Jenger Smith and Mr. Charlie Kerlee in preparing the photoradio-graphs and photomicrographs.
Citation: Journal of the American Animal Hospital Association 38, 6; 10.5326/0380569
Citation: Journal of the American Animal Hospital Association 38, 6; 10.5326/0380569
Representative radiographs from each median sternotomy treatment group. (A) Wire closure lateral. (B) Wire closure dorsoventral. (C) Suture closure lateral. (D) Suture closure dorsoventral.
Photomicrographs of transverse sections of sternal sections from each median sternotomy treatment group. (A) Suture closure. Note the increased fracture gap, increased inflammation, and lack of osteochondral bridging (Hematoxylin and eosin stain, 10×; bar=250 μm). Region within the box is magnified in (B). Again, note the mononuclear inflammation (Hematoxylin and eosin stain, 25×; bar=100 μm). (C) Wire closure. Area within the box shows the region of sternotomy. Complete osteochondral bridging is present in this sample (Hematoxylin and eosin stain, 6.25×; bar=400 μm).
Contributor Notes
