Surg Endosc DOI 10.1007/s00464-014-3926-7
and Other Interventional Techniques
Efficacy of laparoscopic-assisted approach for reversal of Hartmann’s procedure: results from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database Jeanine Arkenbosch • Hiromichi Miyagaki • H. M. C. Shantha Kumara • Xiaohong Yan • Vesna Cekic • Richard L. Whelan Received: 5 April 2014 / Accepted: 1 October 2014 Ó Springer Science+Business Media New York 2014
Abstract Introduction Morbidity after reversal of Hartmann’s procedure remains high. The laparoscopic approach (LAP) may be associated with lower morbidity versus open Hartmann’s closure. This study’s aim is to compare results after LAP and OPEN colostomy takedown and Hartmann’s reversal. Methods The American College of Surgeons National Surgical Quality Improvement Program database was queried from 2005 to 2012 for CPT procedure codes 44227 (LAP) and 44626 (OPEN). Exclusion criteria included: ventilator dependence, ASA class 4 or 5, SIRS, sepsis, emergency case, and advanced malignancy. Demographic parameters were assessed as well as comorbidities and short-term outcomes. Statistical methods used include Fisher’s exact test for categorical variables and Student’s t test for continuous variables.
Presented at the SAGES 2014 Annual Meeting, April 2–5, 2014, Salt Lake City, Utah J. Arkenbosch H. Miyagaki H. M. C. Shantha Kumara X. Yan V. Cekic R. L. Whelan (&) Division of Colon and Rectal Surgery, Department of Surgery, Mount Sinai Roosevelt Hospital, Suite 7B, 425 West, 59th Street, New York, NY 100 19, USA e-mail: [email protected]
Results In total, 4,148 patients underwent stoma closure and Hartmann’s reversal (LAP 732 [17.6 %], OPEN 3,416 [82.3 %]). The mean BMI was lower in the LAP (mean ± SD 27.6 ± 6.6) versus OPEN group (28.3 ± 6.8, p = 0.012). The groups were similar as regards comorbidities except for dyspnea (LAP 5.6 %, OPEN 7.8 %, p = 0.043). The mean surgery times were similar and the median LOS shorter in the LAP versus OPEN groups (5 vs 6 days, p \ 0.0001). A lower overall morbidity rate was noted for the LAP group (18.4 % vs OPEN 27 %, p \ 0.0001) but mortality was statistically similar. Lower rates were noted in the LAP group for the following complications: incisional SSI (10.4 vs 14.1 %, p = 0.033), organ space SSI (3.1 vs 5.0 %, p = 0.033), UTI (1.6 vs 3.3 %, p = 0.005), sepsis (3.4 vs 6.0 %, p = 0.038), and reoperation (3.1 vs 5.4 %, p = 0.011). Conclusion Only 18 % of Hartmann’s reversal’s were done using LAP methods. The LAP and OPEN groups were similar except for gender, BMI, and dyspnea history. LAP methods were associated with a 1 day LOS benefit and significantly lower overall morbidity and lower rates of incisional and deep SSI, UTI, sepsis, and reoperations. Operative length was similar. The short-term results of the LAP approach are superior to the OPEN results. Keywords Open and Laparoscopic-assisted Hartmann’s reversal procedure ACS-NSQIP database Post-operative clinical outcomes
H. M. C. Shantha Kumara e-mail: [email protected] H. Miyagaki Department of Surgery, Saiseika Senri Hospital, 1-1-6 Tsukumodai, Suita, Osaka 565-0862, Japan R. L. Whelan Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA
The Hartmann’s procedure consists of a colorectal resection in conjunction with formation of an end stoma (most often a colostomy) and over sewing of the distal large bowel stump. This procedure, utilized worldwide, is designed to reduce mortality due to anastomotic leakage [1]. It is most often used in patients with left-sided colorectal pathology in whom an
123
Surg Endosc
immediate anastomosis is not feasible and in those judged to be at very high risk for anastomotic leakage. The specific reasons for choosing the Hartmann’s procedure include: (1) emergency surgery for a large bowel perforation or obstruction from any cause, (2) the setting of ischemic bowel, (3) hemodynamic or pulmonary instability, (4) occasionally following an extremely difficult elective sigmoid or rectal resection, and (5) after very low rectal resection for cancer in a patient judged to be a poor candidate for reconnection. Also the lack of a bowel preparation during an urgent or emergent resection is a commonly cited reason for using this approach. The pathologic entities associated with such conditions and circumstances include diverticulitis, cancer, volvulus, ischemic bowel, as well as colonoscopy, colonic stent, and foreign body-related perforations [2, 3]. Restoration of intestinal continuity by reversal of the Hartmann’s procedure is associated with morbidity and mortality rates as high as 50 and 7 %, respectively [4, 5]. At least one-third of the patients undergoing a Hartmann’s procedure will not undergo colostomy closure within a year. Reasons for not reversing the stoma include age, ASA score, high-risk status, patient refusal, and a fear of postoperative complications [4, 6–8]. Patients that refuse or are judged not suitable for Hartmann’s reversal are left with a permanent stoma that affects their quality of life [9]. Lowering the high morbidity and mortality rates associated with Hartmann’s reversal might lead to an increase in the percentage of patients whose colostomies are closed. Laparoscopic reversal of the Hartmann’s procedure was first described in 1993 [10]. Since then a number of retrospective studies have been published that have assessed the minimally invasive approach and compared it to the results following open reversal. These relatively small single institution studies have noted that the minimally invasive approach was associated with a reduced length of stay (LOS) and, in some cases, lower rates of wound infection and anastomotic leakage [4, 11, 12]. Not surprisingly, studies concerning large populations of patients are lacking. In an effort to better evaluate open and closed Hartmann’s reversal methods, a review of the multi-institutional database of the American College of Surgeons National Surgical Quality Improvement Program (ACSNSQIP) was carried out. The aim of this study was to evaluate the effect of laparoscopic-assisted methods in Hartmann’s reversal on postoperative clinical outcomes, compared to results noted with the open approach.
Methods This retrospective review of prospectively gathered data was carried out by searching the ACS-NSQIP database. The ACS-NSQIP is a nationally validated, risk-adjusted,
123
outcome-based program and the purpose of which is to measure and improve the quality of surgical care in the private sector [13]. The ACS-NSQIP program involves over 250 hospitals; data concerning 136 clinical variables are collected, documented, and entered into the database by a clinical nurse reviewer at each site. Amidst the parameters included are variables concerning preoperative risk factors, a variety of intraoperative data points, short-term outcome, and inpatient and outpatient complications that occur within 30 days of surgery [14]. This database categorizes the specific operations performed according to their current procedural terminology (CPT) codes which facilitate the performance of studies concerning a procedure of interest. Only patients registered in the ACS-NSQIP database were discoverable. To identify the patient population for this study, the ACS-NSQIP database was first queried for patients with the international classification of disease 9 (ICD9) codes V55.3 (attention to colostomy) or V44.3 (colostomy) for the 7-year period extending from 2005 to 2012. From the resulting subgroup, only those patients with CPT codes 44426 (closure of enterostomy, large or small intestine, with resection, and colorectal anastomosis e.g., closure of Hartmann type procedure) or 44227 (laparoscopy, surgical, closure of enterostomy, large or small intestine, with resection, and anastomosis) were eligible for this study. The combination of CPT 44626 with the codes 44213 (splenic flexure takedown) or 44180 (lysis of adhesions), in the same patient, was also considered a laparoscopic procedure. Exclusion criteria included: preoperative ventilator dependence, totally dependent functional health status, preoperative septic status (SIRS, septic shock or sepsis), emergency case, advanced malignancy (disseminated cancer or tumor involving CNS), and current pneumonia. Also patients with American Society of Anesthesia (ASA) scores 4 (patients with severe systemic disease that is a constant threat to life), or 5 (moribund patients who are not expected to survive without operation), and those for whom no ASA score was entered were, similarly, excluded from entry into the study population. The ACS-NSQIP data that were considered in this study included: basic demographic data (age, gender, and BMI), the comorbidities of each patient, history of preoperative radiotherapy and/or chemotherapy, ASA score, intraoperative data (operative length, transfusions, etc.), complications, LOS data, and other short-term outcome for the first 30 days after surgery. Complications that occur within 30 days of the operation are captured and included in the NSQIP data regardless of whether they occur in the hospital or as an outpatient. The specific complications that are followed in the NSQIP database and that were considered for this study were superficial and deep surgical site infections (SSI), wound disruption, organ space SSI,
Surg Endosc
urinary tract infection (UTI), renal insufficiency, renal failure, bleeding/transfusion, sepsis, septic shock, reoperation, cardiac arrest and myocardial infarction, pneumonia, unplanned intubation, prolonged intubation, deep vein thrombosis (DVT) requiring therapy, peripheral nerve injury, pulmonary embolus, CVA/stroke with neurological deficit, and death. In a number of instances, in order to simplify the interpretation and analysis of the NSQIP data for this study, several NSQIP complication codes were combined to form a single complication category. For example, the category, incisional SSI, in this study, includes the NSQIP codes for superficial SSI and deep wound SSI. The former is defined as an infection that involves the skin, subcutaneous tissues, whereas the latter refers to SSIs that involve the fascial or muscle layers. The pulmonary complication category includes the NSQIP categories of pneumonia and unplanned intubation. The cardiac complications category includes cardiac arrest and myocardial infarction. DVT, as defined for this study, included DVT requiring therapy and pulmonary embolism. The sepsis category included the NSQIP sepsis and septic shock groupings. The reoperations and deaths that occurred within 30 days of surgery were captured in this data set. The overall complication rate was defined as the sum total of the complications listed in ACSNSQIP database. This retrospective study was approved by the St. Luke’s Roosevelt Institutional Review Board (IRB) and permission was obtained from the American College of Surgeons to use the NSQIP data for research purposes.
Table 1 Characteristics and comorbidities for patients undergoing LAP or OPEN procedures Characteristics
LAP (n = 732)
%
Gender (male)
415
56.7
Age (median)
59
59
NS§
BMI (kg/m2) (mean ± SD)
27.6 ± 6.6
28.3 ± 6.8
0.012§
Results The study population consisted of 4,148 patients enrolled in the NSQIP database between 2005 and 2012 who underwent stoma closure and Hartmann’s reversal as defined by the CPT and ICD9 criteria mentioned above. The laparoscopic group (LAP) contained 732 patients
%
p value
1,792
52.5
0.037*
Diabetes
63
8.6
353
10.3
NS*
Alcohol use
18
2.5
86
2.5
NS*
Dyspnea
41
5.6
267
7.8
0.043*
FHS—partially dependent
18
2.5
61
1.8
NS*
COPD
37
5.1
176
5.2
NS*
Previous percutaneous cardiac intervention
30
4.1
132
3.9
NS*
Previous cardiac surgery
26
3.6
128
3.7
NS*
Hypertension
343
46.9
1,647
48.2
NS*
History of peripheral vascular disease
5
0.7
35
1.0
NS*
Dialysis
9
1.2
33
1.0
NS*
Chronic steroid use
34
4.6
151
4.4
NS*
Weight loss
17
2.3
66
1.9
NS*
Bleeding disorder
22
3.0
85
2.5
NS*
33
4.4
120
3.5
NSà
ASA classification No disturb
Statistics The statistical analyses utilized in this study included the Fisher’s exact tests for categorical variables, the Student’s t test for continuous variables and the Chi-squared test for ASA classifications. All tests were two-tailed and statistical significance was defined as p \ 0.05. All analyses were performed using SPSS version 19 for Windows (IBM Software) SAS Institute, Cary, NC). To adjust for differences in baseline characteristics between the LAP and the OPEN subgroups, we calculated the propensity scores for the two groups of patients and then did propensity score matching (using SPSS version 19).
OPEN (n = 3,416)
Mild disturb
407
55.6
1,854
54.3
Severe disturb
292
39.9
1,442
42.2
BMI body mass index, COPD chronic obstructive pulmonary disease, FHS functional health status, ASA American Society of Anesthesiologists * Fisher’s exact test §
Student t test
à
Chi-square test
[17.6 %], whereas the open group (OPEN) comprised 3,416 patients [82.4 %]. Amidst the study population were 12 patients in whom both the 44626 (open Hartmann’s reversal) and 44227 (laparoscopic Hartmann’s takedown) CPT codes were listed. It was assumed that the operation in these patients was started using minimally invasive methods but was later converted to open methods. According to the ‘‘intent to treat’’ principle, these patients were placed in the laparoscopic group. Table 1 contains the demographic, comorbidity, medical history, and ASA data as well as the percentage of patients who received preoperative radiotherapy, chemotherapy, or transfusions. Table 2 contains the LOS, operative length, complication, and postoperative transfusion data.
123
Surg Endosc Table 2 Operative data and postoperative outcome measures for patients undergoing LAP or OPEN procedures
SD standard deviation, SSI superficial site infection, DVT deep venous thrombosis, UTI urinary tract infection * Fisher’s exact test
Wilcoxon test
§
Student t test
Postoperative outcome measures
LAP (n =732)
Surgery time (mean ± SD)
187.6 ± 86.5
190.4 ± 90.7
NS§
Length of stay (median)
5
6
\0.0001§
OPEN (n =3,416)
%
p value
Length of stay (mean ± SD)
5.4 ± 5.6
Overall complications
135
18.4
921
27.0
\0.0001*
Incisional SSI
76
10.4
483
14.1
0.006*
Organ/space SSI
23
3.1
170
5.0
0.033*
Wound complication
88
12.0
621
18.2
NS*
Pulmonary complications
13
1.8
92
2.7
NS*
Cardiac complications
2
0.3
32
0.9
NS*
DVT
2
0.3
35
1.0
NS*
UTI
12
1.6
113
3.3
0.005*
CVA/stroke
1
0.1
6
0.2
NS*
Bleeding/transfusion
22
3.0
116
3.4
NS*
Septic
25
3.4
204
6.0
0.038*
Reoperation Death
23 2
3.1 0.3
183 19
5.4 0.6
0.011* NS*
The median age for the two groups was the same (59), however, there was a greater proportion of males in the LAP group (56.7 % vs 52.5 [OPEN], p = 0.037). The mean BMI of the LAP patients (mean ± SD 27.6 ± 6.6) was significantly lower than that of the OPEN group (28.3 ± 6.8, p = 0.012). There were no differences in the rate of the various comorbidities except for the category of dyspnea (LAP, 41 patients, 5.6 % vs OPEN, 267, 7.8 %, p = 0.043). There was no significant difference between groups in terms of the ASA category breakdown of patients. As mentioned in the ‘‘Statistics’’ section, to adjust for the above differences in baseline characteristics, propensity score matching was performed. None of the variables were significantly different in propensity score, indicating that there is an equal distribution of confounders between the LAP and OPEN group. These results suggest that it is reasonable to directly compare the LAP and OPEN groups postoperative results and complication rates. The mean surgery time for both groups was not significantly different between the groups (LAP 187.6 ± 86.5 min, OPEN 190.4 ± 90.7, p = 0.442). The LOS was significantly shorter in the LAP (median LOS 5 days) versus the OPEN group (median 6 days, p \ 0.0001). The incidence of overall complications for the LAP procedures (18.4 %) was significantly lower than the morbidity rate in the OPEN group (27.0 %, p \ 0.0001), however, there was no difference in the mortality rates between the groups (LAP 0.3 %, OPEN 0.6 %, p [ 0.556). The incidence of the following complications was found to be significantly lower for the LAP patients versus the OPEN group’s results: incisional SSI (LAP 10.4 % vs OPEN 14.1 %, p = 0.006), organ space SSI (LAP 3.1 %, OPEN 5.0 %, p = 0.033), UTI (LAP 1.6 %,
123
%
7.4 ± 6.4
OPEN 3.3 %, p = 0.005), sepsis (LAP 3.4 %, OPEN 6.0 %, p = 0.038), and reoperations (LAP 3.1 %, OPEN 5.4 %, p = 0.011) (Table 2). There were no significant differences between the LAP and OPEN groups in the occurrence of wound complications [wound disruption] (LAP 12.0 %, OPEN 18.2 %, p = 0.29), pulmonary complications (LAP 1.8 %, OPEN 2.7 %, p = 0.19), cardiac complications (LAP 0.3 %, OPEN 0.9 %, p = 0.073), DVT (LAP 0.3 %, OPEN 1.0 %, p = 0.050), CVA/Stroke (LAP 0.1 %, OPEN 0.2 %, p = 1.000), and bleeding/transfusion (LAP 3.0 %, OPEN 3.4 %, p = 0.651).
Discussion Of the 4,148 Hartmann’s reversal cases found in the ACSNSQIP database for the time period assessed, only 17.6 % were carried out by means of laparoscopic methods. Until recently, previous studies regarded smaller patient populations and were retrospective or case-controlled studies; these investigations have reported that laparoscopic Hartmann’s reversal was associated with lower morbidity and a shorter LOS when compared to equivalent open takedown procedures [11, 12, 15]. Unfortunately, despite the encouraging early results of the smaller studies, the percentage of cases being done using laparoscopic means remain low. In regard to the results of the present study, surprisingly, the mean surgery time did not differ between groups. Importantly, for the LAP group, there was a significant 1 day LOS benefit (median value) as well as a significantly lower incidence of overall complications. In regards to
Surg Endosc
specific complications, in the LAP group, there were significantly fewer incisional SSIs, organ space SSIs (which includes anastomotic leaks and intra abdominal abscesses), UTIs, reoperations, and septic events when compared to the OPEN group. There were no differences between groups in the incidence of wound disruptions, cardiac or pulmonary complications. The patient demographics did not differ significantly between the LAP and OPEN groups, except for a slight preponderance of males and a lower mean BMI in the LAP group. Except for a single comorbidity, dyspnea, there were no significant differences between the LAP and OPEN groups as regards comorbidities. The results of the propensity score matching analysis, mentioned above, suggested that it was reasonable to directly compare the 2 groups. Thus, because of the apparent similarity of the patient groups based on the NSQIP date, it is unlikely that the outcome differences noted, in favor of LAP methods, are attributable to the differences in the patient populations. The noted BMI difference is small (0.7) and not likely influence the results; similarly, the 4 % higher incidence of males in the LAP group is not likely to account for the differences between groups. Overall, these results suggest that the laparoscopic approach is associated with a better outcome, at least up until 30 days post-surgery, than the more commonly performed open Hartmann’s takedown and colostomy closure. The finding of a lower incidence of postoperative complications and shorter LOS in the LAP group is consistent with the results of previous, smaller studies [4, 11, 12]. Unfortunately, the NSQIP database does not include data regarding the conversion rate (to open methods) for the laparoscopic patients [16]. Also it is not possible to determine what percentage of the laparoscopic cases were done using hand-assisted laparoscopic methods, single incision methods, and straight laparoscopic or a laparoscopic-assisted approach. Finally, it is impossible to determine why the LAP or OPEN approach was selected for a given patient. Of note, after the present study had been undertaken and largely completed, the results of a similar study carried out by another group of investigators (Cellini et al.) using the ACS-NSQIP database was published [17]. The study population of the present study (n = 4,148) is 1.61 times the size of the recently published study (n = 2,567). The time period assessed in the first study was 2005–2010, whereas the current study examined the 2005–2012 period. The same CPT codes were utilized to identify the patients although the present study, in addition, required that patients also have one of two ICD9 codes pertaining to ‘‘presence of’’ or ‘‘attention to’’ colostomy. Not surprisingly, the results of the initial ACS-NSQIP study were similar to the current studies findings. Although the assessment of the comorbidity data was done differently (grouping of NSQIP categories) in the two studies, both
concluded that the patient populations were similar. A significantly smaller LOS and a lower rate of major complications as well as wound and septic complications were noted in the initial study. Of note, Cellini et al. did not demonstrate a significant difference in the rate of organ space SSIs as was shown in the current study. Also Cellini et al. did not demonstrate a significantly lower rate of reoperations although a lower percentage of LAP patients (3.3 %) required a trip back to the operating room than in the OPEN group (5.5 %). The present study, with very similar percentages for each group, found a significant difference as regards reoperations most likely due to the larger patient population. Interestingly, in the first paper, 13.1 % of the total population of Hartmann’s closure patients underwent a laparoscopic approach versus 17.6 % in the present study. This difference suggests that, slowly, the utilization rates of LAP methods are increasing.
Conclusion These results suggest that the LAP approach is associated with lower postoperative morbidity and a shorter length of hospitalization and, thus, is preferable to open methods for the closure of Hartmann’s pouches. It also appears that the incidence of cases being done using minimally invasive means in the United States is slowly increasing. Additional studies are needed to determine which specific laparoscopic methods are being used for these cases (handassisted, laparoscopic-assisted, straight laparoscopic, robotic, and single incision approach); they would also, hopefully, reveal how often conversion to laparotomy is required. Future reviews of large databases like the ACSNSQIP will reveal, among other things, whether the percentage of laparoscopic Hartmann’s closure operations being done continues to increase. If this is the case, then the overall incidence of complications would be expected to decrease and the overall LOS shortened. Aknowledgments This study was made possible by a generous donation from Mr. Wade Thompson and family to the Division of Colon and Rectal surgery, Department of Surgery, Mount Sinai Roosevelt Hospital, New York, USA. Disclosures All authors (Jeanine Arkenbosch BS; Hiromichi Miyagaki MD, PhD; Shantha Kumara H.M.C. PhD; Xiaohong Yan PhD; Vesna Cekic RN; Richard L. Whelan MD) have no conflicts of interest or financial ties to disclose.
References 1. Gooszen AW, Tollenaar RA, Geelkerken RH, Smeets HJ, Bemelman WA, VanSchaardenburgh P, Gooszen HG (2001) Prospective study of primary anastomosis following sigmoid
123
Surg Endosc
2. 3. 4.
5.
6.
7.
8.
9.
resection for suspected acute complicated diverticular disease. Br J Surg 88(5):693–697 Bruusgaard C (1947) Volvulus of the sigmoid colon and its treatment. Surgery 22(3):466–478 Boyden AM (1950) The surgical treatment of diverticulitis of the colon. Ann Surg 132(1):94–109 van de Wall BJ, Draaisma WA, Schouten ES, Broeders IA, Consten EC (2010) Conventional and laparoscopic reversal of the Hartmann procedure: a review of literature. J Gastrointest Surg 14(4):743–752 Antolovic D, Reissfelder C, Ozkan T, Galindo L, Bu¨chler MW, Koch M, Weitz J (2011) Restoration of intestinal continuity after Hartmann’s procedure—not a benign operation. Are there predictors for morbidity? Langenbecks Arch Surg 396(7):989–996 Riansuwan W, Hull TL, Millan MM, Hammel JP (2009) Nonreversal of Hartmann’sprocedure for diverticulitis: derivation of a scoring system to predict nonreversal. Dis Colon Rectum 52(8):1400–1408 Banerjee S, Leather AJ, Rennie JA, Samano N, Gonzalez JG, Papagrigoriadis S (2005) Feasibility and morbidity of reversal of Hartmann’s. Colorectal Dis 7(5):454–459 Holland JC, Winter DC, Richardson D (2002) Laparoscopically assisted reversal of Hartmann’s procedure revisited. Surg Laparosc Endosc Percutan Tech 12(4):291–294 Vermeulen J, Gosselink MP, Busschbach JJ, Lange JF (2010) Avoiding or reversing Hartmann’s procedure provides improved
123
10.
11.
12. 13. 14.
15.
16.
17.
quality of life after perforated diverticulitis. J Gastrointest Surg 14(4):651–657 Gorey TF, O’Connell PR, Waldron D, Cronin K, Kerin M, Fitzpatrick JM (1993) Laparoscopically assisted colostomy closure after Hartmann’s procedure. Br J Surg. 80:109 Mazeh H, Greenstein AJ, Swedish K, Nguyen SQ, Lipskar A, Weber KJ, Chin EH, Divino CM (2009) Laparoscopic and open reversal of Hartmann’s procedure. Surg Endosc 23(3):496–502 Golash V (2006) Laparoscopic reversal of Hartmann procedure. J Minim Access Surg 2(4):211–215 American College of Surgeons. http://site.acsnsqip.org/ American College of Surgeons National Surgical Quality Improvement Program. User Guide for the 2007 Participant Use Data File (2008) American College of Surgeons, Chicago Parkin E, Khurshid M, Ravi S, Linn T (2013) Surgical access through the stoma for laparoscopic reversal of Hartmann procedures. Surg Laparosc Endosc Percutan Tech 23(1):41–44 Haughn C, Ju B, Uchal M, Arnaud JP, Reed JF, Bergamaschi R (2008) Complication rates after Hartmann’s reversal. Dis Colon Rectum 51(8):1232–1236 Cellini C, Deeb AP, Sharma A, Monson JR, Fleming FJ (2013) Association between operative approach and complications in patients undergoing Hartmann’s reversal. Br J Surg 100(8): 1094–1099
and Other Interventional Techniques
Efficacy of laparoscopic-assisted approach for reversal of Hartmann’s procedure: results from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database Jeanine Arkenbosch • Hiromichi Miyagaki • H. M. C. Shantha Kumara • Xiaohong Yan • Vesna Cekic • Richard L. Whelan Received: 5 April 2014 / Accepted: 1 October 2014 Ó Springer Science+Business Media New York 2014
Abstract Introduction Morbidity after reversal of Hartmann’s procedure remains high. The laparoscopic approach (LAP) may be associated with lower morbidity versus open Hartmann’s closure. This study’s aim is to compare results after LAP and OPEN colostomy takedown and Hartmann’s reversal. Methods The American College of Surgeons National Surgical Quality Improvement Program database was queried from 2005 to 2012 for CPT procedure codes 44227 (LAP) and 44626 (OPEN). Exclusion criteria included: ventilator dependence, ASA class 4 or 5, SIRS, sepsis, emergency case, and advanced malignancy. Demographic parameters were assessed as well as comorbidities and short-term outcomes. Statistical methods used include Fisher’s exact test for categorical variables and Student’s t test for continuous variables.
Presented at the SAGES 2014 Annual Meeting, April 2–5, 2014, Salt Lake City, Utah J. Arkenbosch H. Miyagaki H. M. C. Shantha Kumara X. Yan V. Cekic R. L. Whelan (&) Division of Colon and Rectal Surgery, Department of Surgery, Mount Sinai Roosevelt Hospital, Suite 7B, 425 West, 59th Street, New York, NY 100 19, USA e-mail: [email protected]
Results In total, 4,148 patients underwent stoma closure and Hartmann’s reversal (LAP 732 [17.6 %], OPEN 3,416 [82.3 %]). The mean BMI was lower in the LAP (mean ± SD 27.6 ± 6.6) versus OPEN group (28.3 ± 6.8, p = 0.012). The groups were similar as regards comorbidities except for dyspnea (LAP 5.6 %, OPEN 7.8 %, p = 0.043). The mean surgery times were similar and the median LOS shorter in the LAP versus OPEN groups (5 vs 6 days, p \ 0.0001). A lower overall morbidity rate was noted for the LAP group (18.4 % vs OPEN 27 %, p \ 0.0001) but mortality was statistically similar. Lower rates were noted in the LAP group for the following complications: incisional SSI (10.4 vs 14.1 %, p = 0.033), organ space SSI (3.1 vs 5.0 %, p = 0.033), UTI (1.6 vs 3.3 %, p = 0.005), sepsis (3.4 vs 6.0 %, p = 0.038), and reoperation (3.1 vs 5.4 %, p = 0.011). Conclusion Only 18 % of Hartmann’s reversal’s were done using LAP methods. The LAP and OPEN groups were similar except for gender, BMI, and dyspnea history. LAP methods were associated with a 1 day LOS benefit and significantly lower overall morbidity and lower rates of incisional and deep SSI, UTI, sepsis, and reoperations. Operative length was similar. The short-term results of the LAP approach are superior to the OPEN results. Keywords Open and Laparoscopic-assisted Hartmann’s reversal procedure ACS-NSQIP database Post-operative clinical outcomes
H. M. C. Shantha Kumara e-mail: [email protected] H. Miyagaki Department of Surgery, Saiseika Senri Hospital, 1-1-6 Tsukumodai, Suita, Osaka 565-0862, Japan R. L. Whelan Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA
The Hartmann’s procedure consists of a colorectal resection in conjunction with formation of an end stoma (most often a colostomy) and over sewing of the distal large bowel stump. This procedure, utilized worldwide, is designed to reduce mortality due to anastomotic leakage [1]. It is most often used in patients with left-sided colorectal pathology in whom an
123
Surg Endosc
immediate anastomosis is not feasible and in those judged to be at very high risk for anastomotic leakage. The specific reasons for choosing the Hartmann’s procedure include: (1) emergency surgery for a large bowel perforation or obstruction from any cause, (2) the setting of ischemic bowel, (3) hemodynamic or pulmonary instability, (4) occasionally following an extremely difficult elective sigmoid or rectal resection, and (5) after very low rectal resection for cancer in a patient judged to be a poor candidate for reconnection. Also the lack of a bowel preparation during an urgent or emergent resection is a commonly cited reason for using this approach. The pathologic entities associated with such conditions and circumstances include diverticulitis, cancer, volvulus, ischemic bowel, as well as colonoscopy, colonic stent, and foreign body-related perforations [2, 3]. Restoration of intestinal continuity by reversal of the Hartmann’s procedure is associated with morbidity and mortality rates as high as 50 and 7 %, respectively [4, 5]. At least one-third of the patients undergoing a Hartmann’s procedure will not undergo colostomy closure within a year. Reasons for not reversing the stoma include age, ASA score, high-risk status, patient refusal, and a fear of postoperative complications [4, 6–8]. Patients that refuse or are judged not suitable for Hartmann’s reversal are left with a permanent stoma that affects their quality of life [9]. Lowering the high morbidity and mortality rates associated with Hartmann’s reversal might lead to an increase in the percentage of patients whose colostomies are closed. Laparoscopic reversal of the Hartmann’s procedure was first described in 1993 [10]. Since then a number of retrospective studies have been published that have assessed the minimally invasive approach and compared it to the results following open reversal. These relatively small single institution studies have noted that the minimally invasive approach was associated with a reduced length of stay (LOS) and, in some cases, lower rates of wound infection and anastomotic leakage [4, 11, 12]. Not surprisingly, studies concerning large populations of patients are lacking. In an effort to better evaluate open and closed Hartmann’s reversal methods, a review of the multi-institutional database of the American College of Surgeons National Surgical Quality Improvement Program (ACSNSQIP) was carried out. The aim of this study was to evaluate the effect of laparoscopic-assisted methods in Hartmann’s reversal on postoperative clinical outcomes, compared to results noted with the open approach.
Methods This retrospective review of prospectively gathered data was carried out by searching the ACS-NSQIP database. The ACS-NSQIP is a nationally validated, risk-adjusted,
123
outcome-based program and the purpose of which is to measure and improve the quality of surgical care in the private sector [13]. The ACS-NSQIP program involves over 250 hospitals; data concerning 136 clinical variables are collected, documented, and entered into the database by a clinical nurse reviewer at each site. Amidst the parameters included are variables concerning preoperative risk factors, a variety of intraoperative data points, short-term outcome, and inpatient and outpatient complications that occur within 30 days of surgery [14]. This database categorizes the specific operations performed according to their current procedural terminology (CPT) codes which facilitate the performance of studies concerning a procedure of interest. Only patients registered in the ACS-NSQIP database were discoverable. To identify the patient population for this study, the ACS-NSQIP database was first queried for patients with the international classification of disease 9 (ICD9) codes V55.3 (attention to colostomy) or V44.3 (colostomy) for the 7-year period extending from 2005 to 2012. From the resulting subgroup, only those patients with CPT codes 44426 (closure of enterostomy, large or small intestine, with resection, and colorectal anastomosis e.g., closure of Hartmann type procedure) or 44227 (laparoscopy, surgical, closure of enterostomy, large or small intestine, with resection, and anastomosis) were eligible for this study. The combination of CPT 44626 with the codes 44213 (splenic flexure takedown) or 44180 (lysis of adhesions), in the same patient, was also considered a laparoscopic procedure. Exclusion criteria included: preoperative ventilator dependence, totally dependent functional health status, preoperative septic status (SIRS, septic shock or sepsis), emergency case, advanced malignancy (disseminated cancer or tumor involving CNS), and current pneumonia. Also patients with American Society of Anesthesia (ASA) scores 4 (patients with severe systemic disease that is a constant threat to life), or 5 (moribund patients who are not expected to survive without operation), and those for whom no ASA score was entered were, similarly, excluded from entry into the study population. The ACS-NSQIP data that were considered in this study included: basic demographic data (age, gender, and BMI), the comorbidities of each patient, history of preoperative radiotherapy and/or chemotherapy, ASA score, intraoperative data (operative length, transfusions, etc.), complications, LOS data, and other short-term outcome for the first 30 days after surgery. Complications that occur within 30 days of the operation are captured and included in the NSQIP data regardless of whether they occur in the hospital or as an outpatient. The specific complications that are followed in the NSQIP database and that were considered for this study were superficial and deep surgical site infections (SSI), wound disruption, organ space SSI,
Surg Endosc
urinary tract infection (UTI), renal insufficiency, renal failure, bleeding/transfusion, sepsis, septic shock, reoperation, cardiac arrest and myocardial infarction, pneumonia, unplanned intubation, prolonged intubation, deep vein thrombosis (DVT) requiring therapy, peripheral nerve injury, pulmonary embolus, CVA/stroke with neurological deficit, and death. In a number of instances, in order to simplify the interpretation and analysis of the NSQIP data for this study, several NSQIP complication codes were combined to form a single complication category. For example, the category, incisional SSI, in this study, includes the NSQIP codes for superficial SSI and deep wound SSI. The former is defined as an infection that involves the skin, subcutaneous tissues, whereas the latter refers to SSIs that involve the fascial or muscle layers. The pulmonary complication category includes the NSQIP categories of pneumonia and unplanned intubation. The cardiac complications category includes cardiac arrest and myocardial infarction. DVT, as defined for this study, included DVT requiring therapy and pulmonary embolism. The sepsis category included the NSQIP sepsis and septic shock groupings. The reoperations and deaths that occurred within 30 days of surgery were captured in this data set. The overall complication rate was defined as the sum total of the complications listed in ACSNSQIP database. This retrospective study was approved by the St. Luke’s Roosevelt Institutional Review Board (IRB) and permission was obtained from the American College of Surgeons to use the NSQIP data for research purposes.
Table 1 Characteristics and comorbidities for patients undergoing LAP or OPEN procedures Characteristics
LAP (n = 732)
%
Gender (male)
415
56.7
Age (median)
59
59
NS§
BMI (kg/m2) (mean ± SD)
27.6 ± 6.6
28.3 ± 6.8
0.012§
Results The study population consisted of 4,148 patients enrolled in the NSQIP database between 2005 and 2012 who underwent stoma closure and Hartmann’s reversal as defined by the CPT and ICD9 criteria mentioned above. The laparoscopic group (LAP) contained 732 patients
%
p value
1,792
52.5
0.037*
Diabetes
63
8.6
353
10.3
NS*
Alcohol use
18
2.5
86
2.5
NS*
Dyspnea
41
5.6
267
7.8
0.043*
FHS—partially dependent
18
2.5
61
1.8
NS*
COPD
37
5.1
176
5.2
NS*
Previous percutaneous cardiac intervention
30
4.1
132
3.9
NS*
Previous cardiac surgery
26
3.6
128
3.7
NS*
Hypertension
343
46.9
1,647
48.2
NS*
History of peripheral vascular disease
5
0.7
35
1.0
NS*
Dialysis
9
1.2
33
1.0
NS*
Chronic steroid use
34
4.6
151
4.4
NS*
Weight loss
17
2.3
66
1.9
NS*
Bleeding disorder
22
3.0
85
2.5
NS*
33
4.4
120
3.5
NSà
ASA classification No disturb
Statistics The statistical analyses utilized in this study included the Fisher’s exact tests for categorical variables, the Student’s t test for continuous variables and the Chi-squared test for ASA classifications. All tests were two-tailed and statistical significance was defined as p \ 0.05. All analyses were performed using SPSS version 19 for Windows (IBM Software) SAS Institute, Cary, NC). To adjust for differences in baseline characteristics between the LAP and the OPEN subgroups, we calculated the propensity scores for the two groups of patients and then did propensity score matching (using SPSS version 19).
OPEN (n = 3,416)
Mild disturb
407
55.6
1,854
54.3
Severe disturb
292
39.9
1,442
42.2
BMI body mass index, COPD chronic obstructive pulmonary disease, FHS functional health status, ASA American Society of Anesthesiologists * Fisher’s exact test §
Student t test
à
Chi-square test
[17.6 %], whereas the open group (OPEN) comprised 3,416 patients [82.4 %]. Amidst the study population were 12 patients in whom both the 44626 (open Hartmann’s reversal) and 44227 (laparoscopic Hartmann’s takedown) CPT codes were listed. It was assumed that the operation in these patients was started using minimally invasive methods but was later converted to open methods. According to the ‘‘intent to treat’’ principle, these patients were placed in the laparoscopic group. Table 1 contains the demographic, comorbidity, medical history, and ASA data as well as the percentage of patients who received preoperative radiotherapy, chemotherapy, or transfusions. Table 2 contains the LOS, operative length, complication, and postoperative transfusion data.
123
Surg Endosc Table 2 Operative data and postoperative outcome measures for patients undergoing LAP or OPEN procedures
SD standard deviation, SSI superficial site infection, DVT deep venous thrombosis, UTI urinary tract infection * Fisher’s exact test
Wilcoxon test
§
Student t test
Postoperative outcome measures
LAP (n =732)
Surgery time (mean ± SD)
187.6 ± 86.5
190.4 ± 90.7
NS§
Length of stay (median)
5
6
\0.0001§
OPEN (n =3,416)
%
p value
Length of stay (mean ± SD)
5.4 ± 5.6
Overall complications
135
18.4
921
27.0
\0.0001*
Incisional SSI
76
10.4
483
14.1
0.006*
Organ/space SSI
23
3.1
170
5.0
0.033*
Wound complication
88
12.0
621
18.2
NS*
Pulmonary complications
13
1.8
92
2.7
NS*
Cardiac complications
2
0.3
32
0.9
NS*
DVT
2
0.3
35
1.0
NS*
UTI
12
1.6
113
3.3
0.005*
CVA/stroke
1
0.1
6
0.2
NS*
Bleeding/transfusion
22
3.0
116
3.4
NS*
Septic
25
3.4
204
6.0
0.038*
Reoperation Death
23 2
3.1 0.3
183 19
5.4 0.6
0.011* NS*
The median age for the two groups was the same (59), however, there was a greater proportion of males in the LAP group (56.7 % vs 52.5 [OPEN], p = 0.037). The mean BMI of the LAP patients (mean ± SD 27.6 ± 6.6) was significantly lower than that of the OPEN group (28.3 ± 6.8, p = 0.012). There were no differences in the rate of the various comorbidities except for the category of dyspnea (LAP, 41 patients, 5.6 % vs OPEN, 267, 7.8 %, p = 0.043). There was no significant difference between groups in terms of the ASA category breakdown of patients. As mentioned in the ‘‘Statistics’’ section, to adjust for the above differences in baseline characteristics, propensity score matching was performed. None of the variables were significantly different in propensity score, indicating that there is an equal distribution of confounders between the LAP and OPEN group. These results suggest that it is reasonable to directly compare the LAP and OPEN groups postoperative results and complication rates. The mean surgery time for both groups was not significantly different between the groups (LAP 187.6 ± 86.5 min, OPEN 190.4 ± 90.7, p = 0.442). The LOS was significantly shorter in the LAP (median LOS 5 days) versus the OPEN group (median 6 days, p \ 0.0001). The incidence of overall complications for the LAP procedures (18.4 %) was significantly lower than the morbidity rate in the OPEN group (27.0 %, p \ 0.0001), however, there was no difference in the mortality rates between the groups (LAP 0.3 %, OPEN 0.6 %, p [ 0.556). The incidence of the following complications was found to be significantly lower for the LAP patients versus the OPEN group’s results: incisional SSI (LAP 10.4 % vs OPEN 14.1 %, p = 0.006), organ space SSI (LAP 3.1 %, OPEN 5.0 %, p = 0.033), UTI (LAP 1.6 %,
123
%
7.4 ± 6.4
OPEN 3.3 %, p = 0.005), sepsis (LAP 3.4 %, OPEN 6.0 %, p = 0.038), and reoperations (LAP 3.1 %, OPEN 5.4 %, p = 0.011) (Table 2). There were no significant differences between the LAP and OPEN groups in the occurrence of wound complications [wound disruption] (LAP 12.0 %, OPEN 18.2 %, p = 0.29), pulmonary complications (LAP 1.8 %, OPEN 2.7 %, p = 0.19), cardiac complications (LAP 0.3 %, OPEN 0.9 %, p = 0.073), DVT (LAP 0.3 %, OPEN 1.0 %, p = 0.050), CVA/Stroke (LAP 0.1 %, OPEN 0.2 %, p = 1.000), and bleeding/transfusion (LAP 3.0 %, OPEN 3.4 %, p = 0.651).
Discussion Of the 4,148 Hartmann’s reversal cases found in the ACSNSQIP database for the time period assessed, only 17.6 % were carried out by means of laparoscopic methods. Until recently, previous studies regarded smaller patient populations and were retrospective or case-controlled studies; these investigations have reported that laparoscopic Hartmann’s reversal was associated with lower morbidity and a shorter LOS when compared to equivalent open takedown procedures [11, 12, 15]. Unfortunately, despite the encouraging early results of the smaller studies, the percentage of cases being done using laparoscopic means remain low. In regard to the results of the present study, surprisingly, the mean surgery time did not differ between groups. Importantly, for the LAP group, there was a significant 1 day LOS benefit (median value) as well as a significantly lower incidence of overall complications. In regards to
Surg Endosc
specific complications, in the LAP group, there were significantly fewer incisional SSIs, organ space SSIs (which includes anastomotic leaks and intra abdominal abscesses), UTIs, reoperations, and septic events when compared to the OPEN group. There were no differences between groups in the incidence of wound disruptions, cardiac or pulmonary complications. The patient demographics did not differ significantly between the LAP and OPEN groups, except for a slight preponderance of males and a lower mean BMI in the LAP group. Except for a single comorbidity, dyspnea, there were no significant differences between the LAP and OPEN groups as regards comorbidities. The results of the propensity score matching analysis, mentioned above, suggested that it was reasonable to directly compare the 2 groups. Thus, because of the apparent similarity of the patient groups based on the NSQIP date, it is unlikely that the outcome differences noted, in favor of LAP methods, are attributable to the differences in the patient populations. The noted BMI difference is small (0.7) and not likely influence the results; similarly, the 4 % higher incidence of males in the LAP group is not likely to account for the differences between groups. Overall, these results suggest that the laparoscopic approach is associated with a better outcome, at least up until 30 days post-surgery, than the more commonly performed open Hartmann’s takedown and colostomy closure. The finding of a lower incidence of postoperative complications and shorter LOS in the LAP group is consistent with the results of previous, smaller studies [4, 11, 12]. Unfortunately, the NSQIP database does not include data regarding the conversion rate (to open methods) for the laparoscopic patients [16]. Also it is not possible to determine what percentage of the laparoscopic cases were done using hand-assisted laparoscopic methods, single incision methods, and straight laparoscopic or a laparoscopic-assisted approach. Finally, it is impossible to determine why the LAP or OPEN approach was selected for a given patient. Of note, after the present study had been undertaken and largely completed, the results of a similar study carried out by another group of investigators (Cellini et al.) using the ACS-NSQIP database was published [17]. The study population of the present study (n = 4,148) is 1.61 times the size of the recently published study (n = 2,567). The time period assessed in the first study was 2005–2010, whereas the current study examined the 2005–2012 period. The same CPT codes were utilized to identify the patients although the present study, in addition, required that patients also have one of two ICD9 codes pertaining to ‘‘presence of’’ or ‘‘attention to’’ colostomy. Not surprisingly, the results of the initial ACS-NSQIP study were similar to the current studies findings. Although the assessment of the comorbidity data was done differently (grouping of NSQIP categories) in the two studies, both
concluded that the patient populations were similar. A significantly smaller LOS and a lower rate of major complications as well as wound and septic complications were noted in the initial study. Of note, Cellini et al. did not demonstrate a significant difference in the rate of organ space SSIs as was shown in the current study. Also Cellini et al. did not demonstrate a significantly lower rate of reoperations although a lower percentage of LAP patients (3.3 %) required a trip back to the operating room than in the OPEN group (5.5 %). The present study, with very similar percentages for each group, found a significant difference as regards reoperations most likely due to the larger patient population. Interestingly, in the first paper, 13.1 % of the total population of Hartmann’s closure patients underwent a laparoscopic approach versus 17.6 % in the present study. This difference suggests that, slowly, the utilization rates of LAP methods are increasing.
Conclusion These results suggest that the LAP approach is associated with lower postoperative morbidity and a shorter length of hospitalization and, thus, is preferable to open methods for the closure of Hartmann’s pouches. It also appears that the incidence of cases being done using minimally invasive means in the United States is slowly increasing. Additional studies are needed to determine which specific laparoscopic methods are being used for these cases (handassisted, laparoscopic-assisted, straight laparoscopic, robotic, and single incision approach); they would also, hopefully, reveal how often conversion to laparotomy is required. Future reviews of large databases like the ACSNSQIP will reveal, among other things, whether the percentage of laparoscopic Hartmann’s closure operations being done continues to increase. If this is the case, then the overall incidence of complications would be expected to decrease and the overall LOS shortened. Aknowledgments This study was made possible by a generous donation from Mr. Wade Thompson and family to the Division of Colon and Rectal surgery, Department of Surgery, Mount Sinai Roosevelt Hospital, New York, USA. Disclosures All authors (Jeanine Arkenbosch BS; Hiromichi Miyagaki MD, PhD; Shantha Kumara H.M.C. PhD; Xiaohong Yan PhD; Vesna Cekic RN; Richard L. Whelan MD) have no conflicts of interest or financial ties to disclose.
References 1. Gooszen AW, Tollenaar RA, Geelkerken RH, Smeets HJ, Bemelman WA, VanSchaardenburgh P, Gooszen HG (2001) Prospective study of primary anastomosis following sigmoid
123
Surg Endosc
2. 3. 4.
5.
6.
7.
8.
9.
resection for suspected acute complicated diverticular disease. Br J Surg 88(5):693–697 Bruusgaard C (1947) Volvulus of the sigmoid colon and its treatment. Surgery 22(3):466–478 Boyden AM (1950) The surgical treatment of diverticulitis of the colon. Ann Surg 132(1):94–109 van de Wall BJ, Draaisma WA, Schouten ES, Broeders IA, Consten EC (2010) Conventional and laparoscopic reversal of the Hartmann procedure: a review of literature. J Gastrointest Surg 14(4):743–752 Antolovic D, Reissfelder C, Ozkan T, Galindo L, Bu¨chler MW, Koch M, Weitz J (2011) Restoration of intestinal continuity after Hartmann’s procedure—not a benign operation. Are there predictors for morbidity? Langenbecks Arch Surg 396(7):989–996 Riansuwan W, Hull TL, Millan MM, Hammel JP (2009) Nonreversal of Hartmann’sprocedure for diverticulitis: derivation of a scoring system to predict nonreversal. Dis Colon Rectum 52(8):1400–1408 Banerjee S, Leather AJ, Rennie JA, Samano N, Gonzalez JG, Papagrigoriadis S (2005) Feasibility and morbidity of reversal of Hartmann’s. Colorectal Dis 7(5):454–459 Holland JC, Winter DC, Richardson D (2002) Laparoscopically assisted reversal of Hartmann’s procedure revisited. Surg Laparosc Endosc Percutan Tech 12(4):291–294 Vermeulen J, Gosselink MP, Busschbach JJ, Lange JF (2010) Avoiding or reversing Hartmann’s procedure provides improved
123
10.
11.
12. 13. 14.
15.
16.
17.
quality of life after perforated diverticulitis. J Gastrointest Surg 14(4):651–657 Gorey TF, O’Connell PR, Waldron D, Cronin K, Kerin M, Fitzpatrick JM (1993) Laparoscopically assisted colostomy closure after Hartmann’s procedure. Br J Surg. 80:109 Mazeh H, Greenstein AJ, Swedish K, Nguyen SQ, Lipskar A, Weber KJ, Chin EH, Divino CM (2009) Laparoscopic and open reversal of Hartmann’s procedure. Surg Endosc 23(3):496–502 Golash V (2006) Laparoscopic reversal of Hartmann procedure. J Minim Access Surg 2(4):211–215 American College of Surgeons. http://site.acsnsqip.org/ American College of Surgeons National Surgical Quality Improvement Program. User Guide for the 2007 Participant Use Data File (2008) American College of Surgeons, Chicago Parkin E, Khurshid M, Ravi S, Linn T (2013) Surgical access through the stoma for laparoscopic reversal of Hartmann procedures. Surg Laparosc Endosc Percutan Tech 23(1):41–44 Haughn C, Ju B, Uchal M, Arnaud JP, Reed JF, Bergamaschi R (2008) Complication rates after Hartmann’s reversal. Dis Colon Rectum 51(8):1232–1236 Cellini C, Deeb AP, Sharma A, Monson JR, Fleming FJ (2013) Association between operative approach and complications in patients undergoing Hartmann’s reversal. Br J Surg 100(8): 1094–1099
