Surgical Endoscopy

, Volume 24, Issue 7, pp 1727–1736

Peritoneal inflammatory response of natural orifice translumenal endoscopic surgery (NOTES) versus laparoscopy with carbon dioxide and air pneumoperitoneum

Authors

    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Michael F. McGee
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Leandro T. Cavazzola
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Steve Schomisch
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Mehrdad Nikfarjam
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Jessica Bailey
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Tripurari Mishra
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Benjamin K. Poulose
    • Department of SurgeryVanderbilt University Medical Center, Medical Center North
  • Young-Joon Lee
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Jeffrey L. Ponsky
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
  • Jeffrey M. Marks
    • Department of SurgeryUniversity Hospitals Case Medical Center, Case Western Reserve University School of Medicine
Article

DOI: 10.1007/s00464-009-0839-y

Cite this article as:
Trunzo, J.A., McGee, M.F., Cavazzola, L.T. et al. Surg Endosc (2010) 24: 1727. doi:10.1007/s00464-009-0839-y
  • 93 Views

Abstract

Background

The immunologic and physiologic effects of natural orifice translumenal endoscopic surgery (NOTES) versus traditional surgical approaches are poorly understood. Previous investigations have shown that NOTES and laparoscopy share similar inflammatory cytokine profiles except for a possible late-phase tissue necrosis factor-α (TNF-α) depression with NOTES. The local peritoneal reaction and immunomodulatory influence of pneumoperitoneum agents in NOTES also are not known and may play an important role in altering the physiologic insult induced by NOTES.

Methods

In this study, 51 animals were divided into four study groups, which respectively underwent abdominal exploration via transgastric NOTES using room air (AIR) or carbon dioxide (CO2) or via laparoscopy (LX) using AIR or CO2 for pneumoperitoneum. Laparotomy and sham surgeries were additionally performed as control conditions. Measurements of TNF-α, interleukin-1β (IL-1β), and IL-6 were performed for peritoneal fluid collected after 0, 2, 4, and 6 h and on postoperative days (PODs) 1, 2, and 7.

Results

Of the 45 animals assessed, 6 were excluded because of technical operative complications. The findings showed that LX-CO2 generated the most pronounced response with all three inflammatory markers. However, no significant differences were detected between LX-CO2 and either NOTES group at these peak points. No differences were encountered between NOTES-CO2 and NOTES-AIR. Subgroup comparisons showed significantly higher levels of TNF-α and IL-6 with NOTES-CO2 than with LX-AIR on POD 1 (p = 0.022) and POD 2 (p = 0.002). The LX-CO2 subgroup had significantly higher levels of TNF-α than the LX-AIR subgroup at 4 h (p = 0.013) and on POD 1 (p = 0.021). No late-phase TNF-α depression occurred in the NOTES animals.

Conclusion

The local inflammatory reaction to NOTES was similar to that with traditional laparoscopy, and the previously described late-phase systemic TNF-α depression in serum was not reproduced. At the peritoneal level, NOTES is no more physiologically stressful than laparoscopy. Furthermore, regardless of which gas was used, the role of the pneumoperitoneum agent did not affect the cytokine profile after NOTES, suggesting that air pneumoperitoneum is adequate for NOTES.

Keywords

EndoscopylaparoscopyImmune system

The implementation of natural orifice translumenal endoscopy surgery (NOTES) has challenged researchers, surgeons, and gastroenterologists alike. Many areas of continued uncertainty with this approach have hampered its progress and have demanded more scientific proof to validate its use. Major concerns about visceral organ leak, infectious risk, and the physiologic effect of enteric contamination continue to demand further scrutiny.

Early in the overall investigation of NOTES, our group began to examine the question of its physiologic impact. Our initial porcine model investigated the serum cytokine activity produced when these animals were subjected to NOTES, laparoscopy, and laparotomy [1]. This study’s major conclusions were that the initial physiologic response was very similar among the groups but that in the late postoperative period, NOTES generated an immunosuppressed response. However, a recent study reexamining this conclusion did not produce similar results [2]. The findings showed a slight continued increase in inflammation over time. Consequently, an understanding of the physiologic impact of NOTES still remains unclear.

One factor not addressed by either aforementioned study is the role the insufflation gas plays in the physiological response after NOTES versus laparoscopy. The immunomodulatory effect of carbon dioxide (CO2) previously demonstrated in laparoscopy [3, 4] may augment this response and ultimately provide a potential benefit with NOTES. Furthermore, special attention should be given to the local peritoneal reaction as opposed to monitoring only proinflammatory mediators in the serum.

Several studies have demonstrated that peritoneal fluid harbors a much greater concentration of cytokine markers [57] and therefore may provide a more accurate assessment of the local inflammatory response. Thus, the primary goal of this investigation was to assess the physiologic effect of NOTES compared with laparoscopy at the local peritoneal level while correlating the systemic response with serum cortisol. Additionally, an evaluation for any potential benefit in implementing CO2 with NOTES was performed.

Methods

Animals

This experiment was completed after approval and supervision from the Case Western Reserve University Institutional Animal Care and Use Committee. For the study, 51 female domestic farm pigs weighing 40 kg were obtained from a USDA-approved vendor (Pine View Farms, Valley City, OH, USA). These pigs underwent a 7-day quarantine and acclimation period at the authors’ institution. During this period, veterinary personnel evaluated each animal to ensure its baseline health. All the animals were subjected to the same husbandry procedures. The animals were caged individually, fed the same diet, and allowed unlimited access to water. The climate-controlled animal facility consisted of 12-h light-dark cycles that mimicked the animals’ native environment.

Groups

The animals were assigned to four study groups: NOTES peritoneoscopy with room air (AIR), NOTES peritoneoscopy with CO2, exploratory laparoscopy (LX) with CO2, and LX with AIR. The control groups included open laparotomy (OPEN) for positive control and a sham procedure (SHAM) for negative control. To eliminate potential variability between the study groups, preoperative, anesthesia, and postoperative care protocols were identical for all the groups.

Preoperative care and anesthesia

All the animals were restricted from food 24 h before the scheduled date of surgery or sham procedure but were allowed to drink water ad libitum. The animals were first sedated with 10 mg/kg of intramuscular tiletamine HCl and zolazepam (Telazol; Fort Dodge, Animal Health, Fort Dodge, IA, USA) to facilitate transfer to the operating suite and allow for intubation.

After sedation, 4 ml of blood was obtained through venipuncture of the internal jugular vein at time 0 h (0HR). Endotracheal intubation then was performed with a 6-mm endotracheal tube, and the animals were mechanically ventilated at 12 respirations/min with a tidal volume of 15 ml/kg and a fraction of inspired oxygen (FiO2) of 100%. Inhaled isoflurane (AErrane; Baxter Healthcare Corp., Deerfield, IL, USA) was administered at 1.5% for exactly 90 min while the animal underwent the assigned surgery or sham procedure.

Enrofloxacin 2.5 mg/kg (Baytril; Bayer Healthcare LLC, Shawnee Mission, KA, USA) was administered once intramuscularly for all the groups preoperatively. Hair was removed from the anterior abdominal wall of all the animals, and the skin was prepped with 2% chlorhexidine gluconate and 70% isopropyl alcohol (Chloraprep; Cardinal Health, Leawood, KS, USA). Each animal was draped with sterile towels and sheets, exposing the mouth, right anterior neck, and abdominal wall. All the animals then underwent placement of an abdominal drain and right internal jugular venous catheter for specimen collection.

Placement of the abdominal drain

Using the Hasson technique, a flat 10-mm Jackson-Pratt abdominal drain was inserted into the peritoneal cavity 2 cm below the umbilicus and directed toward the pelvis. A direct aspiration of peritoneal fluid (3–5 ml) was collected immediately through the drain (at 0HR) and stored for analysis. The drain tubing was tunneled subcutaneously to exit the skin in a more protected location, near the right flank, and then sealed with a stopcock.

Next, the point of entry was closed in two layers: the fascia was closed with a #0 glycolide/lactide copolymer suture (Polysorb; Syneture, Norwalk, CT, USA) and the skin with interrupted 4-0 nylon suture (Monosof; Norwalk, CT, USA). The externalized drain tubing was secured to the skin with 4-0 Monosof suture.

Placement of the right internal jugular venous catheter

Simultaneously with placement of the abdominal drain, a second surgeon performed a cut-down to access the right internal jugular vein through a 5-cm vertical incision. Once the vein was identified, a central venous catheter (Hickman; CR Bard Inc., Murray Hill, NJ, USA) was inserted through a venotomy and advanced 10 cm toward the chest. The proximal vein was ligated, and the distal vein was looped using 3-0 silk ties (Sofsilk; Syneture) to secure the catheter within the vein. The catheter was tunneled to the lateral neck and brought through the skin. The incision was closed with two layers using a running 3-0 Polysorb to close the platysma and an interrupted closure of 4-0 Monosof for the skin. The catheter was aspirated and then flushed with heparin until the next blood draw. The external portion of the catheter was secured to the skin using 4-0 Monosof.

Surgery

Once the abdominal drain and venous catheter were placed, the animals underwent their assigned treatment. For the animals assigned to sham procedures, no further interventions were performed, and they were subsequently recovered after the standardized 90 min of anesthesia were completed. All the laparotomy and laparoscopic equipment was sterile. The NOTES equipment underwent high-grade disinfection with 0.55% orthophthalaldehyde solution (Cidex OPA Solution; Advanced Sterilization Products, Irvine, CA, USA) between animals.

The LX animals underwent a standardized three-port abdominal exploration. A 15-mm trocar was placed at the umbilicus (superior to the drain position) using the Hassan technique. Pneumoperitoneum was created with a laparoscopic insufflator to 15 mmHg through the trocar.

For the LX-AIR group, the laparoscopic insufflator was modified to use air for inflation. Two additional 5-mm ports were placed bilaterally along the linea semilunaris under laparoscopic visualization. Laparoscopic bowel graspers then were inserted to assist with the standardized 10-min abdominal exploration. At completion of the surgery, the trocars were withdrawn, and the fascial umbilical defect was closed with a figure eight using #0 Polysorb suture. The overlying skin was closed for all the ports using 4-0 Monosof in an interrupted fashion.

The NOTES animals underwent nonsterile, standardized, transgastric NOTES access using a previously described technique [8]. Briefly, a standard percutaneous endoscopic gastrostomy (PEG)-type technique was performed to identify an anterior location for transgastric entry. The gastrotomy was created using an endoscopic needleknife sphincterotome (Boston Scientific Corp., Natick, MA, USA) and an 18–20-mm esophageal dilating balloon (Boston Scientific) at the chosen point. The endoscope then was inserted through the gastrotomy. Once the endoscope was intraperitoneal, a 10-min standardized abdominal exploration was performed using endoscopic biopsy forceps for tissue manipulation. Abdominal insufflation for both NOTES groups was unregulated for consistency.

The NOTES-AIR group was managed with a standard endoscopy using visualization as the cue for insufflating air. Insufflation for the NOTES-CO2 group was provided by attaching the tubing of a laparoscopic insufflator to a 14-gauge angiocatheter and inserting it into a working channel of the endoscope. Next, CO2 was insufflated until the visualization was adequate, then removed, with abdominal pressures left unknown to the operator. If during the procedure, additional insufflation was required, the angiocatheter was again reinserted briefly.

The NOTES gastrotomy finally was closed with an endoscopic tissue plicating device (NDO Plicator; NDO Surgical, Mansfield, MA, USA) using a technique developed and reported by our group [9, 10]. Briefly, two or three pledgeted suture implants were deployed by the NDO Plicator to invert the tissue and close the gastrotomy with full-thickness bites.

The OPEN animals underwent the same standard abdominal exploration via midline laparotomy. Electrocautery was used to create a standardized 25-cm-long midline laparotomy extending from the xiphoid process inferior. After manual exploration, each animal was closed in layers. A running looped #1 glycolide and trimetheylene carbonate suture (Maxon; Syneture) was used to close the fascia, and the overlying skin was closed with 4-0 Monosof in an interrupted fashion. All animal incisions were covered with a protective dressing (Tegaderm; 3 M, St. Paul, MN, USA).

Anesthesia recovery, postoperative observation, and specimen collection

All the animals followed the same recovery protocol. After 90 min of anesthesia, inhaled isoflurane was stopped, and the animals were extubated once adequate respiration returned. Repeated blood draws and direct aspirations of peritoneal fluid were performed at hours 2 (2HR), 4 (4HR), and 6 (6HR) after the initiation of anesthesia. Subsequent collections occurred on PODs 1, 2, and 7. For pain control, a prophylactic single dose of buprenorphine (5 μg/kg) opioid was injected intramuscularly at the 2HR collection. The animals were permitted to eat and drink immediately after surgery and were observed by a team of veterinary and surgical personnel daily until necropsy.

Necropsy

On the POD7, the animals were sedated using 10 mg/kg of tiletamine HCl and zolazepam, and then the final sample collection was taken (POD7). Exploratory laparotomy was performed to rule out incidental pathology.

Specimen analysis

Blood samples were analyzed by an independent laboratory (University Hospitals Case Medical Center Core Laboratory, Cleveland, OH, USA) for serum cortisol level testing. Peritoneal fluid was frozen and stored at −80°C until the time of analysis. Concentrations of tissue necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and IL-6 were measured in the peritoneal fluid using a quantitative “sandwich” enzyme-linked immunosorbent assay (ELISA) kit (R & D System, Minneapolis, MN, USA) according to the manufacturer’s instructions.

Samples were dispensed into 96-well microtiter plates containing an immobilized monoclonal antibody specific for the relevant porcine cytokine. After incubation, unbound protein was rinsed from the wells, to which an enzyme-linked antibody directed against the relevant cytokine was added. After further rinsing to remove unbound enzyme-linked antibody, a substrate solution that reacts with the enzyme to produce a blue color was added. Assays were performed on duplicate samples. Cytokine concentrations were determined spectrophotometrically at 450 nmol/l with a plate correction of 540 nmol/l by comparison with a standard curve using a four-parameter logistic curve fit.

Statistical analysis

Initially, the SHAM and OPEN groups were compared head-to-head with the aim to validate the use of peritoneal fluid as a model for comparing surgical stress from a minimally invasive procedure with that of a maximally invasive procedure. Next, the four experimental groups were analyzed together as a study group cohort. The animals that incurred major technical operative or major non–surgery-related complications that would significantly alter the immune response were excluded. Non–surgery-related complications, such as pneumonia, were confirmed by the in-house veterinary staff at necropsy.

Before analysis, each animal’s 0HR cytokine and cortisol level was used as a baseline for that individual animal. This 0HR level then was normalized to zero, and the difference was applied at all collection points for that animal. This was performed to limit animal-to-animal variability and to create a uniform baseline for comparison. Mean values then were calculated for the groups at each time point and expressed as mean ± standard error of the mean (SEM). Student’s t-test was applied when the OPEN and SHAM groups were compared, and an analysis of variance (ANOVA) was performed for study group comparisons. An alpha level of 0.05 was used as the level of significance. The statistical software SPSS version 11.5 (SPSS, Inc., Chicago, IL, USA) was used for analysis.

Results

The 51 animals were separated into the following six groups: SHAM (n = 5), OPEN (n = 5), NOTES-AIR (n = 11), NOTES-CO2 (n = 10), LX-CO2 (n = 10), and LX-AIR (n = 10). Because of major complications deemed to have a significant impact on immune response, 6 of the 51 animals were excluded from the study (Table 1). The remaining 45 animals were analyzed.
Table 1

Major complication exclusions

Group

Excluded patients (n)

Complication

SHAM

1

Iatrogenic enterotomy requiring repair

OPEN

0

 

NOTES-AIR

2

Iatrogenic colon perforation requiring repair

Developed severe rectal prolapse postoperativelya

NOTES-CO2

2

Severe sepsis secondary to gastric leak

Bilateral pneumoniab

LX-AIR

0

 

LX-CO2

1

Bilateral pneumoniac

SHAM procedure, OPEN procedure, NOTES natural orifice translumenal surgery, NOTES-AIR NOTES with room air, NOTES-CO2 NOTES with carbon dioxide, LX-AIR laparoscopy with room air, LS-CO2 laparoscopy with carbon dioxide

aOccurrence from unknown etiology

bRespiratory failure and failure to thrive requiring euthanasia on postoperative day 4

cRespiratory failure resulting in spontaneous death on postoperative day 1

Control subjects

Comparison of the control subjects (SHAM vs OPEN) on the day of surgery showed a statistically significant difference in peak elevation at the 4HR time point for TNF-α (25 ± 6 vs. 523 ± 179 pg/ml, respectively; p = 0.05) (Fig. 1) and cortisol (4.2 ± 2.7 vs. 14.2 ± 6.1 μg/dl, respectively; p = 0.045) (Fig. 2). A difference also was seen in the IL-6 peak at 6HR, with the SHAM group measuring only 1.3 ± 0.3 ng/ml and the OPEN group reaching 29.2 ± 4.2 ng/ml (p = 0.003) (Fig. 3). There were no significant differences identified for IL-1β at any time point (Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig1_HTML.gif
Fig. 1

Comparison of the TNF-α control groups. The statistical difference is denoted by the asterisk (*) at the indicated time points

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig2_HTML.gif
Fig. 2

Comparison of the serum cortisol control groups. The statistical difference is denoted by the asterisk (*) at the indicated time points

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig3_HTML.gif
Fig. 3

Comparison of the IL-6 control groups. The statistical difference denoted by the asterisk (*) at the indicated time points

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig4_HTML.gif
Fig. 4

Comparison of the IL-1 control groups

On POD7, a spike in inflammation was demonstrated by all markers in the SHAM group, although this was not statistically significant compared with the OPEN group. The three cytokine levels increased at least 20-fold from the POD2 measurements recorded. At necropsy, it was noted that 2 of 4 SHAM and 2 of 5 OPEN animals displayed evidence of gross inflammatory reaction, such as erythema or fibrinous exudates, along the resting position of the abdominal drain at necropsy.

Study groups

A study group comparison showed that LX-CO2 generated the most pronounced mean concentration of all cytokine markers tested, nearly equivalent to the mean peak cortisol level of NOTES-CO2. Subgroup comparisons, however, showed no difference in LX-CO2 at the inflammatory peaks of either NOTES group. However, a statistical difference in peak level was observed between LX-CO2 and LX-AIR at 4HR for TNF-α (3,746 ± 935 vs. 1,307 ± 217 pg/ml, respectively; p = 0.013) (Fig. 5) and at 6HR for IL-6 (12.5 ± 1.7 vs. 7.6 ± 0.9 ng/ml, respectively; p = 0.031) (Fig. 6). Statistically, LX-CO2 further remained greater, with TNF-α measuring 98 ± 35 pg/ml compared with LX-AIR, which measured only 1.8 ± 15 pg/ml on POD1 (p = 0.021). This pattern continued on POD2, with IL-6 measuring 0.5 ± 0.2 ng/ml compared with 0.1 ± 0.03 ng/ml (p = 0.044), respectively. The NOTES-CO2 levels similarly remained elevated over those of LX-AIR at those time points again with both TNF-α (p = 0.022) and IL-6 (p = 0.002). As shown in Fig. 7, IL-1β displayed no statistical differences between groups.
https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig5_HTML.gif
Fig. 5

Comparison of the TNF-α study group cohort. The statistical difference in subgroup analysis is denoted by the asterisk (*) at the indicated time points

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig6_HTML.gif
Fig. 6

Comparison of the IL-6 study group cohort. The statistical difference in the subgroup analysis is denoted by the asterisk (*) at the indicated time points

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig7_HTML.gif
Fig. 7

Comparison of the IL-1 study group cohort

During the early inflammatory phase at 2HR, cortisol data showed that NOTES-CO2 generated a significantly greater initial rise in stress compared with LX-CO2 (p = 0.013), approaching significance over LX-AIR (p = 0.051) (Fig. 8). In terms of cortisol, NOTES-AIR also was significantly greater than LX-CO2 at 2HR (p = 0.047). With IL-6, the only early phase (2HR) comparison that showed a difference was that NOTES-AIR was slightly greater than NOTES-CO2 (p = 0.036).
https://static-content.springer.com/image/art%3A10.1007%2Fs00464-009-0839-y/MediaObjects/464_2009_839_Fig8_HTML.gif
Fig. 8

Comparison of the serum cortisol study group cohort. The statistical difference in subgroup analysis is denoted by the asterisk (*) at the indicated time points

During the final late postoperative phase, an inflammatory elevation was noted again on POD7 for all four study groups excluding IL-1β. However, the only statistical difference identified was solely with IL-6, for which NOTES-AIR (5.3 ± 2.3 ng/ml) was significantly greater than both LX-CO2 (0.5 ± 0.3 ng/ml; p = 0.021) and LX-AIR (0.8 ± 0.5 ng/ml; p = 0.024). Drain-related complications were again observed in the study groups at necropsy (Table 2).
Table 2

Peri-drain inflammatory reaction and/or fibrinous exudates present at necropsy

Group

Adverse events (n)

SHAM

2

OPEN

2

NOTES-AIR

4

NOTES-CO2

2

LX-AIR

1

LX-CO2

0

SHAM procedure, OPEN procedure, NOTES natural orifice translumenal surgery, NOTES-AIR NOTES with room air, NOTES-CO2 NOTES with carbon dioxide, LX-AIR laparoscopy with room air, LS-CO2 laparoscopy with carbon dioxide

Discussion

The natural inflammatory response to an abdominal intervention was demonstrated by a series of acute-phase reactants and proinflammatory mediators. Macrophages play a major role in the local inflammatory reaction that occurs after an invasion of the peritoneal cavity. This response maintains an important role in the recruitment of white blood cells to an area of injury or infection by facilitating a cascade of wound-healing and infection-fighting reactions. With NOTES, the degree to which this occurs has not been well characterized.

The NOTES technique is speculated to be less invasive than laparoscopy due to the absence of transabdominal incisions. However, this approach subjects the patient to the contaminants of the intraluminal cavity, which is nearly impossible to sterilize. The evident concern is that this may result in an excessive source for infection and inflammatory reaction. Furthermore, it could make NOTES more physiologically stressful than laparoscopy despite the fact that it is “incisionless.”

The other confounding factor is the type of pneumoperitoneum gas used at the time of surgery. Traditional flexible endoscopy uses room air as the source of insufflation. Findings have clearly shown that CO2 generates an acidic environment in laparoscopy, modulating the overall inflammatory response [3, 11, 12]. Some have suggested that this physiologic effect is more beneficial than that of air pneumoperitoneum [13, 14], which may provide a mode for improving the overall physiologic and inflammatory profile of NOTES. In this study, we attempted a reproduction of this effect to determine whether CO2 insufflation in NOTES could offer an overall benefit.

In designing our model, we aimed to obtain a more sensitive gauge of the inflammatory response by observing cytokine concentrations locally in the peritoneal fluid. In our previous work [1], we analyzed acute-phase protein markers of inflammation in serum samples and found very low or undetectable concentrations. Prior animal [5, 6] and human [7] studies comparing laparotomy with laparoscopy have demonstrated much higher concentrations of inflammatory markers in peritoneal fluid than the levels routinely reported in serum after surgical stress [15]. Therefore, in this follow-up study, we began by comparing a positive control (OPEN) and a negative control (SHAM) to validate our model for collecting peritoneal fluid. The intention was to quantitate the values that can be obtained from maximal and minimal stress and additionally to demonstrate our ability to identify a difference between them.

We observed a statistically greater peak response in the OPEN group than in the SHAM group, which supported our model design. Furthermore, the local inflammatory response we observed was 10–4,000-fold greater in concentration than reported for the control groups of our previous work [1]. This confirmed the utility of collecting peritoneal fluid to monitor the stress response in this porcine model and may further pose an improved standard for peritoneal stress monitoring in future investigations.

The subsequent evaluation of the study groups produced only one consistent finding across the proinflammatory markers, with LX-CO2 generating the highest inflammatory peak response overall. However, the only peaks for which we observed a statistical difference were in subgroup comparisons of LX-CO2 to LX-AIR with TNF-α at 4HR and IL-6 at 6HR. At these time points, the NOTES groups showed no statistical difference from LX-CO2 or from each other.

One concern with proinflammatory marker collections is whether the peak response has been captured during serial testing. For TNF-α, the peak occurred by 4 h and began sloping down at the 6-h mark, suggesting that we nearly captured its maximal response.

The IL-6 data conflicted with the data of a porcine study by Ure et al. [14], who also compared CO2 laparoscopy with room air laparoscopy using peritoneal fluid. Their findings showed that room air generated a statistically greater IL-6 response than CO2 at the 2-h interval. However, no other samples were obtained after the 2-h collection until 48 h later, which may have left a large window for continued inflammatory response to occur.

Gebhard et al. [16] observed hourly levels of IL-6 for 12 h after stress from trauma and found that patients rated with the highest injury severity score (>32) peaked at 6 h, whereas those with a score lower than nine peaked at 2 h. Their data suggest that IL-6 levels need to be followed out to at least 6 h for the most stressful events to capture the maximal response, which is consistent with our observations.

We therefore could assume that we have accurately monitored the peaks for TNF-α and IL-6 in our study groups. From this, we concluded that, regardless of the insufflation gas used, NOTES generated no more physiologic stress at the inflammatory peak than traditional laparoscopy, as assessed by the markers chosen for this study. When we then broke down the inflammatory reaction into the initial rise, the peak, the early postoperative phase (POD1 and POD2), and the late postoperative phase (POD7), we found that the serum cortisol levels were useful during the initial rise. Cortisol previously was demonstrated to provide an early indication of increased physiologic stress [17]. It has been identified further as a sensitive predictor of stress when observed in animal models [18, 19].

In our observations, we found no difference in serum cortisol levels between groups except in the early inflammatory phase during the period of initial rise. At this point, both NOTES groups generated a more rapid response than the LX-CO2 group before eventually peaking together between 4 and 6 h. This indicates a more immediate stress reaction than generated by NOTES independently of insufflation gas. However, it did not result in an overall greater peak response on the initial operative day or a further display of any differences between groups during the early or late postoperative phases. It is unclear whether a more rapidly generated stress response would have an impact on eventual outcomes when the overall response is essentially similar.

One additional confounding variable that may have altered the interpretation of these results was the “actual” procedure time compared with the standardized anesthesia time used in this study. Anesthesia time was chosen because it can be easily controlled and standardized across all groups. However, we acknowledge that our NOTES closure was of longer duration than the closure of the laparoscopic port sites, which could have influenced our comparative results. Measuring the time the animals were under direct stress from tissue manipulation may have provided some additional information about our observations, especially during the initial stress phase.

Another aim of this investigation was to determine whether the insufflation gas used played a role in modulating the inflammatory reaction during the early and late postoperative phases after NOTES. Much of the previous literature describing the effect of CO2 with respect to laparoscopy has suggested an overall attenuated inflammatory response [3, 11, 12, 20] that may further provide an immunologic benefit [3, 4]. The acidic environment created has been the main contributing factor believed to facilitate this physiologic result. Conversely, “room air” laparoscopy has been shown to generate a greater inflammatory response [13, 14].

In our investigation, however, we observed that the CO2 groups almost universally generated a greater inflammatory reaction than the air groups overall. We did not observe an attenuated inflammatory response in the CO2 groups by comparison. This was not only evident at the main peak levels, as discussed earlier, but also was present during the early postoperative phase, in which both CO2 groups maintained higher TNF-α levels on POD1 and then continued with higher IL-6 levels on POD2. Again, during this phase, we could not demonstrate a benefit in implementing NOTES with CO2.

In our prior investigation, in serum, we observed late-phase immunosuppression with TNF-α for a traditional “room air” NOTES group. In the current study, we were not able to reproduce this finding in peritoneal fluid because an unexpected spike in inflammation was noted instead across both the control and study groups on POD7. The most concerning was the dramatic increase seen in the SHAM group. This we could only speculate, together with the necropsy findings, was a consequence of a secondary inflammatory insult attributable to the chronic indwelling abdominal drain.

Furthermore, as shown in Table 2, there was clear evidence of inflammatory reaction across groups to the drains grossly detected at necropsy. Placement of intraabdominal drains has been demonstrated repeatedly to increase surgical wound infections [2123], especially when left more than 3 days [24]. However, true intraabdominal infections specifically have not shown a well-documented association. This may be due to the fact that abdominal drains are routinely managed to allow continuous drainage and thus to avoid infectious collections.

This was not the case with our model, in which the drain was clamped between sample collections. As a result of this complication, we cannot truly validate our results on POD7. We do, however, recognize that both NOTES groups had higher mean levels of IL-6 and cortisol than the laparoscopic counterparts on that day. This could suggest that the bacterial contamination we generally anticipate from a NOTES procedure may have spurred a late-phase response. Again, we can only speculate that the initial peritoneal contamination anticipated from NOTES may have seeded the drain at the time of surgery.

Furthermore, because the drain was not freely evacuating fluid as would be routine, it perhaps began to develop a peri-drain infection much greater than seen in other groups. Continuing survival past the 7-day period may have evolved into a frank peri-drain abscess. From this result alone, we caution the placement of a permanent foreign body postoperatively in a NOTES procedure, which intuitively will remain contaminated at completion of the procedure. In future studies, removal of the drain on POD2 and collection of the POD7 specimen at necropsy could reduce the incidence of this complication.

Another limitation we acknowledge was the exclusion of six animals due to technical and non–surgical-related complications. From a technical standpoint, two bowel injuries occurred during the surgical intervention as well as a single gastric leak. These were thought to affect grossly the peritoneal inflammatory response and inappropriately skew the analysis.

Leak with the transgastric approach has repeatedly shown varying results depending on the type of closure performed and the technique used, although all published studies used relatively small cohorts mainly in animal models. This further underscores the importance of continued investigation of this key NOTES component. However, this study was not designed to test the physiologic impact of leak, so including the ill animal with a leak would have falsely elevated the NOTES group based on a technical complication.

In addition, two animals were excluded due to severe respiratory compromise caused by pneumonia, resulting in early death, as verified by the in-house veterinary staff. Pneumonia is a major drawback to performing physiologic studies in porcine models due to its well-described incidence in pigs that live in close quarters [25, 26]. Although unavoidable, this complication resulted in early deaths, which were not consistent with the amount of stress incurred by the remaining cohort.

Finally, a single severe rectal prolapse occurred, which also is a known complication associated with pigs [27]. The physiologic impact that this would have produced from the peritoneum is truly unknown. However, this complication still was thought not to represent fairly the other animals in the group, warranting the animal’s exclusion.

It is important to recognize that this study had a small sample size and used an animal model. A larger group for comparison may have shown other statistical differences that we could not produce in this evaluation. However, even with the small groups studied, the pattern of inflammatory reaction was fairly uniform and consistent between the control and study animals in terms of rise and fall.

We do acknowledge, however, that this animal study does not prove the safety of NOTES. The physiologic effect demonstrated in this study may not be exactly equivalent to what could be seen in a human trial and can be proven only by continued investigation. Furthermore, this study had numerous technical limitations and, as shown by our drain complication, other potential postoperative risks need to be addressed on a larger scale before the safety of NOTES can be proven.

In conclusion, the maximum local peritoneal inflammatory reaction of NOTES is similar to that of traditional laparoscopy. However, the previously described late-phase systemic TNF-α depression could not be reproduced at the peritoneal level. Complications with the chronic indwelling drain may have masked the late-phase effect. Nevertheless, regardless which gas is applied, the role of the pneumoperitoneum agent used in NOTES did not affect the cytokine profile compared with the gold standard of CO2 laparoscopy. At the peritoneal level, NOTES was shown to be no more physiologically stressful locally than laparoscopy at the point of peak response and further during the early postoperative period.

From an immunologic perspective, CO2 pneumoperitoneum did not convey immunomodulatory effects with NOTES, so air pneumoperitoneum may be adequate for application in NOTES procedures. However, it still is important to recognize that additional clinical concerns associated with room air insufflations, such as the effects of slower resorption leading to prolonged pain and overall risk of air embolis, have yet to be examined with NOTES and thus require further investigation in the future.

Disclosures

Joseph A. Trunzo, Michael F. McGee, Leandro T. Cavazzola, Steve Schomisch, Mehrdad Nikfarjam, Jessica Bailey, Tripurari Mishra, Benjamin K. Poulose, Young-Joon Lee, Jeffrey L. Ponsky, and Jeffrey M. Marks have no conflict of interests or financial ties to disclose.

Copyright information

© Springer Science+Business Media, LLC 2010