Mini-laparotomy with Adjunctive Care versus Laparoscopy for Placement of Gastric Electrical Stimulation

Abstract

We compared outcomes for two gastric electrical stimulation placement strategies, mini-laparotomy with adjunctive care (MLAC) versus laparoscopy without adjunctive care (LAPA). For electrode placement, the peritoneal cavity was accessed with either a single 2.5 to 3.0 cm midline incision (MLAC) or three trocar incisions (LAPA). For both groups, generator was placed subcutaneously over the anterior rectus sheath. For MLAC, adjunctive pain control measures were used for placement of both electrode and generator (transversus abdominus plane block). For LAPA, those that could not be completed by laparoscopy were converted to traditional open approach and kept in the analysis. MLAC (n = 128) resulted in shorter operative times than LAPA (n = 37) (median operative time: 87.5 vs 137.0 minutes, P ≤ 0.01). Hospital length of stay was also shorter for MLAC than for LAPA (median: 2.0 vs 3.0 days, P ≤ 0.01) without any increase in readmission rates to the hospital within 30 days of discharge (11.0 vs 16.2%, P = 0.39). After equalizing learning curves, these differences were even greater (median operative time: 84.5 vs 137.0 minutes, P < 0.01; median length of stay: 1.0 vs 3.0 days; P < 0.01) without increasing 30-day readmission rates (9.1 vs 16.2%, P = 0.25). For implantation of gastric electrical stimulators, mini-laparotomy can result in improved outcomes when coupled with adjunctive pain control measures.

Gastric electrical stimulation (GES) has been used extensively for the treatment of patients with the symptoms of gastroparesis,^1–4 but little data exist on short-term outcome such as postoperative hospital lengths of stay (LOS). We have previously shown that hospital LOS for GES implantation varies with type of surgery and perioperative care.⁵ We now compare outcomes for GES via minilaparotomy with adjunctive care (MLAC) versus laparoscopy without adjunctive care (LAPA) to measure equivalency.

Methods

Patient Selection

Patients were evaluated with baseline standardized solid gastric emptying tests⁶ and underwent pre-operative temporary endoscopic GES in both groups before permanent implant. Only patients responsive to temporary endoscopic GES, which included assessment of symptoms, gastric emptying, and gastric electrical activity,⁷ underwent permanent GES placement.⁸ The permanent GES system includes intramuscular electrodes and a battery pack with device and received United States Food and Drug Administration approval as a Humanitarian Use Device in 2000.^{9, 10}

Adjunctive Pain Control Measures

Two different methods were used to locally control pain and thereby limit need for systemic narcotics. Limiting the need for narcotics was the primary goal of these measures as narcotics may worsen gastroparesis and limit the response to GES. To control pain associated with generator placement, transversus abdominus plane (TAP) block was performed before incision was made with bupivacaine liposome (Exparel, Pacira Pharmaceuticals, Parsippany, NJ). To control pain associated with transperitoneal incisions, a pain control system (OnQ Pain Buster PostOp Pain Relief System, I-Flow LLC, Irvine, CA) was placed in the preperitoneum to either side of the wound. The system was loaded with 750 cc of 0.2 per cent ropivacaine and each of the two catheters was set to run at 5 cc/h. The system was removed when empty, typically by patients after discharge. TAP block could not be used alone as it would not cover the subxiphoid laparotomy incision and OnQ could not be used alone as the catheters were not long enough to cover both incisions.

Laparoscopy without Adjunctive Care

Both strategies for GES placement represent single surgeon experiences. The first cohort of patients was primarily approached with laparoscopy with conversion to open laparotomy if the laparoscopic approach was not feasible. No adjunctive pain control measures were used. Three trocars were used, with one lengthened at the skin level for generator placement.

Minilaparotomy with Adjunctive Care

The second cohort of patients followed the first as the program transitioned from one surgeon (RC) to another (MH). Patients with MLAC had preoperative TAP block to cover the area of generator placement as the operatively placed pain system (OnQ) could not cover both areas. Laparotomy was performed through a 2.5 to 3.0 cm midline incision and lengthened only if needed for adhesiolysis. If the ribs were widely spaced away from midline, then the incision was made just below the xiphoid process. If the ribs were narrowly spaced close to midline, then the incision was made up to 3 cm below the xiphoid process. Retraction was provided by disposable wound protractor (Alexis Wound Protractor/Retractor (small), Applied Medical, Rancho Santa Margarita, CA) that typically lengthened the wound to approximately 4 to 5 cm (Fig. 1). A long 5-mm trocar was used to tunnel the electrodes from the subcutaneous periumbilical pocket to the wound protractor opening.

Fig. 1.

(A) Wound protractor in place and electrodes tunneled using a 5-mm laparoscopic trocar. (B) Incisions closed after protractor removal and pain system placement.

Additional Measures

Both operative approaches allowed for the recording of gastric electrical activity using both the permanent and temporary electrodes. The use of two electrodes intraoperatively allows for low-resolution gastric mapping.

Outcome Measures

Health-related quality of life was measured by investigator-derived independent outcome measure score (IDIOMS)¹¹ at baseline. Follow-up gastrointestinal (GI) symptoms were recorded at baseline. Results were reported as median or mean ± standard deviation and were compared by Wilcoxon rank-sum test or Student’s t test, respectively. LOS was reported as median because means were not normally distributed.

Results

Patient Characteristics

For the first cohort of patients (LAPA), 188 underwent surgery by a single surgeon from April 2002 to April 2013, of these, 151 cases were performed at another institution and were not available for this analysis. The learning curve for this surgeon was therefore not considered. Of the remaining 37 cases that were used for this analysis, 4 were converted to open. For the second cohort of patients (MLAC), 128 patients underwent surgery by a second surgeon from June 2013 to December 2014. All these cases were performed at the same institution as the 37 LAPA cases. Therefore, the learning curve was included, unlike for the LAPA group. The two groups were similar in age (44 ± 14 years old for MLAC vs 42 ± 14 years old for LAPA, P =0.40), gender (81.2% female for MLAC vs 91.9% female for LAPA, P =0.43), and baseline GI total symptom scores (15.2 ± 3.4 for MLAC vs 14.8 ± 4.1 for LAPA, P = 0.54). The etiologies of gastroparesis were: idiopathic (MLAC 64.0% and LAPA 45.9%), diabetic (MLAC 23.5% and LAPA 29.8%), and postsurgical (MLAC 12.5% and LAPA 24.3%).

Outcomes Data

MLAC cases were shorter than LAPA cases (median operative time: 87.5 vs 137.0 minutes, P ≤ 0.01) (Fig. 2A). As the surgeon placed the pain control system, this additional procedure was accounted for in the operative time. The TAP block performed by the anesthesiologist did not significantly prolong anesthesia only time (total anesthesia time minus operative time, median: 45.5 vs 38.0 minutes; P =0.14). Hospital LOS was significantly shorter for MLAC than for LAPA (median: 2.0 vs 3.0 days; P ≤ 0.01) (Fig. 2B) without any increase in readmission rates to the hospital within 30 days of discharge (11.0 vs 16.2%; P = 0.39).

Fig. 2.

(A) Median operative time and (B) LOS for MLAC (learning curve included) and LAPA groups (left) and MLAC-LCC (learning-curve-censored) and LAPA groups (right). For both comparisons, the LAPA group did not include surgeon learning curve.

Learning-curve-censured Outcomes Data

As the LAPA surgeon’s learning curve likely occurred during the 151 cases performed at his prior institution and was therefore not accounted for in the analysis, the MLAC surgeon’s learning curve was characterized for both operative time (Fig. 3A) and LOS (Fig. 3B). As the decreasing trend in median operative time and LOS appeared to stabilize after 40 cases, these cases were excluded and are reported as learning-curve-censured data (MLAC-LCC). After the first 40 cases, there was also less variability in operative time and LOS.

Fig. 3.

Learning curve associated with MLAC for both (A) operative time and (B) LOS.

With the learning-curve-censured data, operative time and LOS were reduced even further for MLAC-LCC compared with LAPA [operative time, median: 84.5 vs 137.0 minutes, P ≤ 0.01 (Fig. 2A); LOS, median: 1.0 vs 3.0 days, P ≤ 0.01 (Fig. 2B)]. Again, no differences in anesthesia only time was observed (median: 44.5 vs 38.0 minutes; P = 0.13). Readmission (30 day) rates for MLAC-LCC were again statistically similar in this group compared to LAPA (median: 9.1 vs 16.2%; P = 0.25) though the MLAC-LCC readmission rate decreased even further.

Baseline GI symptoms (vomiting, nausea, anorexia, bloating, and pain) were equivalent in the two groups (P > 0.05); however, there were statistically significant differences in baseline IDIOMS scoring (Table 1). In the MLAC group, patients had lower baseline IDIOMS (MLAC 19.7 ± 3.5, LAPA 21.2 ± 3.7; P = 0.028) and severity of illness (MLAC 7.8 ± 1.2, LAPA 8.4 ± 1.3; P ≤ 0.01) scores, both of which represent lower degree of severity. In the learning-curve-censured group, there were also lower baseline IDIOMS (MLAC-LCC 19.3 ± 3.1; P ≤ 0.01) and severity of illness (MLAC-LCC 7.4 ± 1.1; P ≤ 0.01) scores, as well as lower baseline intensity of service scores (MLAC-LCC 6.4 ± 1.9, LAPA 7.2 ± 2.4; P = 0.05). From the available data, there were 11 IDIOMS scores missing from the MLAC group.

Table 1.

Patient Demographics, Baseline Symptoms, and Baseline Quality of Life Scoring

	MLAC	MLAC-LCC	LAPA	P Value
				MLAC vs LAPA	MLAC-LCC vs LAPA
Age	44 ± 14	44 ± 15	42 ± 14.2	0.40	0.35
Gender (% female)	81.2	79.5	91.9	0.43	0.09
Vomiting	2.6 ± 1.4	2.3 ± 1.4	2.4 ± 1.4	0.61	0.97
Nausea	3.3 ± 0.9	3.0 ± 0.9	3.4 ± 1.0	0.60	0.54
Anorexia	3.1 ± 1.1	2.9 ± 1.1	3.3 ± 1.0	0.32	0.35
Bloating	3.0 ± 1.2	2.9 ± 1.2	2.9 ± 1.3	0.40	0.66
Pain	3.2 ± 1.0	3.1 ± 1.1	2.9 ± 1.4	0.08	0.15
Total symptom score	15.2 ± 3.4	14.2 ± 3.3	14.8 ± 4.1	0.54	0.83
Severity of illness	7.8 ± 1.2	7.4 ± 1.1	8.4 ± 1.3	0.01	<0.01
Organ systems involved	5.4 ± 1.2	5.4 ± 1.2	5.5 ± 1.0	0.62	0.65
Intensity of service	6.5 ± 2.2	6.4 ± 1.9	7.2 ± 2.4	0.09	0.05
Total IDIOMS score	19.7 ± 3.5	19.3 ± 3.1	21.2 ± 3.7	0.03	<0.01

Each symptom score has a maximum of 5 with maximum total symptom score of 20. Each quality of life score (IDIOMS) has a maximum of 10 with maximum total IDIOMS of 30.

Discussion

The current study shows that the strategy of MLAC is a reasonable alternative to LAPA. MLAC has been shown here to shorten operative times and LOS without increasing readmission rates. Although these two operative approaches represent contrasting practices of two surgeons, it unlikely that this work is showing surgeon-specific differences: both approaches are straightforward with relatively short operative times. Any surgeon, particularly one accustomed to caring for chronically ill patients, could likely adopt either approach. These adjunctive measures are unlikely to impact laparoscopic cases (though not studied) as trocar sites were too far apart to be effectively covered by the pain control system.

Although the learning curve presented here appeared to end at about 40 cases, several technical points can be made to help shorten this for others. First, midline incision placement was varied based on angle of the patient’s costal margin. If the costal margin is steep, then the left lobe of the liver will block access to the antrum when the incision is made immediately below the xiphoid. Second, when tunneling the trocar through the abdominal wall, the fascia was pierced in the midline to avoid penetrating the rectus muscle, which can help avoid several weeks to months of abdominal wall spasm. Third, close attention was paid to the location of the “soaker hose” portion of pain control system (OnQ) catheters. The catheters should lay directly lateral to the wound on either side, as placement too far distally will result in increased pain. Also, pulling the catheters out too proximally will result in leakage from the exit site rather than to the wound even if sealed in place with a tissue adhesive. Fourth, intravenous non-narcotic pain medications were preferentially used (e.g., ketorolac and acetaminophen) as narcotics exacerbate gastroparesis. Fifth, the skin incision for generator placement was placed transversely midway between costal margin and the anterior superior iliac spine. The device was secured at its inferior margin allowing for rotation away from the ribs when the patient flexes at the hip.

Although the use of laparoscopy versus traditional surgical approaches⁵ showed that laparoscopy shortened postoperative LOS, these data show that mini-laparotomy with ancillary measures is shorter still than laparoscopy. The reason for this shorter LOS is most likely attributable to better control of postoperative pain. Although the placement of GES devices can in some cases help with the long-term treatment of abdominal pain,¹² both groups should have benefitted equally from the placement of GES devices.

The placement of temporary GES devices, pre-operatively, can help with the selection of patients who may respond to permanent GES.^{13, 14} In this study, both groups had preoperative placement of a temporary device, and thus should have responded equally. Although the mechanism of GES is still debated, one mechanism may be that of normalization of gastric electrical dysrhythmias.^{15, 16} As with other factors, the possible effects of GES on gastric electrical abnormalities should be the same in each group.

The demographics and etiology of the patients’ illness were similar between both groups at baseline. There was a small but statistically significant difference in health-related quality of life between the two groups at baseline with the patients who underwent minilaparotomy with ancillary treatments being slightly less severe. However, all patients were quite significantly symptomatic and had complex illnesses¹⁷ and thus the underlying illnesses appear unlikely to explain the differences in postoperative LOS.¹⁸

Our next step will be to selectively perform these cases on an outpatient basis. We have found that non-obese, opioid naïve patients with no prior gastric surgery may not need an overnight admission; however, some do. Therefore, adequate safety measures such as a hotel nearby and a clinic visit the next morning are essential for caring for such chronically ill patients. Furthermore, setting consistent and reasonable expectations for the patients facilitates their adherence to the treatment plan since chronically ill patients may have decreased emotional reserve and may exhibit dysfunctional behavior if there is little structure to the program. Although the operation is not usually technically challenging, patient care before and after the procedure can be demanding on both patients and providers.

Conclusion

For implantation of gastric electrical stimulators in this group of patients with gastroparesis, minilaparotomy is as effective in improving symptoms as laparoscopic implantation and can result in shorter LOS when coupled with adjunctive measures to limit incision size and relieve pain.