Investigation of the Indications for Conservative Therapy to Treat Perforated Gastroduodenal Ulcers

Transcription

1 85 Original Article J. St. Marianna Univ. Vol. 7, pp , 2016 Investigation of the Indications for Conservative Therapy to Treat Perforated Gastroduodenal Ulcers Kazuya Niwa, Shinya Mikami, Takeharu Enomoto, and Takehito Otsubo (Received for Publication: August 22, 2016) Abstract At our hospital, we have been using the criteria reported by Watanabe et al. to determine the treatment plan for patients with of perforated gastroduodenal ulcers to determine whether they should undergo surgery. Between April 2003 and March 2015, we evaluated 117 cases of perforation, and among these, we selected conservative treatment for 14 cases. One of these 14 cases required conversion to surgery. Although we selected surgical treatment in 103 cases, at surgery, the perforation was already covered in 19 cases. The Watanabe score was found to have a sensitivity of 0.99 and specificity of 0.41, and none of the cases judged using this scoring method had perforation as a cause of death. We have therefore determined the validity of the Watanabe score. However, we believe that we performed an excessive amount of surgery as we found some cases in which the perforation was already covered and therefore could have been managed conservatively. To avoid this, we have proposed the following new scoring system, which uses coefficients to provide weighting during logistic regression analysis for the four items found to have a significant difference in terms of their Watanabe scores. The formula is calculated as: [if symptoms of peritoneal irritation are limited to the epigastrium] [if no severe concomitant disease] [if duration is less than 6 hours from the time of onset to initial examination] [if progression of ascites is limited to the epigastrium]. If the total score is 2.08 or greater, then conservative treatment is selected. This new scoring system not only decreases the number of items evaluated when compared to the Watanabe system, but with a sensitivity of 0.94 and specificity of 0.63, it may also provide a more definitive method for determining the treatment plan. Key words perforated gastroduodenal ulcer, conservative treatment Introduction In conjunction with the advancements in pharmaceutical therapy for peptic ulcers in recent years, numerous reports have stated that selection of conservative treatment for the early treatment of upper gastrointestinal perforations was effective 55% of the time 1 5). The 2015Japanese Society of Gastroenterology Guidelines for the Management of Peptic Ulcers 6) state that the indications for surgical treatment include 1 a long clinical course after onset, 2 peritonitis not limited to the epigastric region, 3 a large volume of ascites, 4 a large volume of gastric contents, 5 patients aged 70 years, 6 severe concomitant diseases, and 7 patients with unstable hemodynamics. However, the level of evidence is low, there is ambiguity, and as there are still ongoing discussions regarding clear scoring for indications, no consensus has been reached. Since 2006, we have been performing treatment in accordance with the independent scoring system reported by Watanabe et al. 7) When we reviewed the cases treated surgically, we could confirm that the surgical findings indicated that the perforation in some of the cases had already been covered by the time of surgery, by the greater omentum for example, whereas in some cases, the perfora Department of Gastrointestinal and General Surgery, St. Marianna University School of Medicine 55

2 86 Niwa K Mikami S et al tion was small and had simply closed. These may have been cases that could have been treated by means of conservative treatment. In the present study, we retrospectively evaluated the blood test findings, imaging findings, and surgical findings in cases of gastroduodenal perforated ulcers and re-evaluated the validity of the Watanabe score with the aim of establishing more definitive selection criteria. I. Subjects and Methods Fig. 1. Treatment algorithm for upper gastrointestinal tract perforation SIRS: systemic inflammatory response syndrome; WBC: white blood cells. Our subjects were 117 patients with perforated gastroduodenal ulcers who were admitted to and treated at the Department of Gastrointestinal and General Surgery, St. Marianna University School of Medicine in the 13 years between April 2003 and March Patients with iatrogenic perforation were excluded from the study. Below are the selection criteria reported by Watanabe et al. that are used to determine the method for treating these perforated gastroduodenal ulcers (Fig. 1). 1) Subjects meeting the criteria for systemic inflammatory response syndrome (SIRS) at the time of the initial examination were diagnosed as being in poor general condition and were operated on immediately. Subjects met the criteria for SIRS if they scored 2 or more on items (1) (4) as follows (each item below is worth 1 point): (1) Body temperature: > 38 C or < 36 C (2) Heart rate: > 90 beats/min (3) Respiratory rate:>20 beats/min or PaCO 2 < 32 mmhg (4) White blood cell count: > 12,000 cells/μl, < 4,000 cells/μl, or >10% immature leukocytes 2) If the subjects did not meet the criteria for SIRS, then subjects achieving a Watanabe score of 3 or greater were treated conservatively, and those with a Watanabe score of 2 or less were operated on immediately (each item below is worth 1 point). A) Duration of less than 5 hours from the time of onset to initial examination. B) The absence of a full stomach (as examined on CT images). C) Abdominal findings: The symptoms of peritoneal irritation are limited to the epigastrium. D) Progression of ascites is limited to the epigastrium (as examined on CT images). E) No severe concomitant disease. Re-evaluation was performed approximately every 24 hours but was performed every 12 hours in elderly patients aged 70 years or older. On the basis of this diagnostic plan, we investigated the subsequent clinical course in the group that underwent conservative treatment and assigned them to the conservative group (C group). We then investigated which of the subjects underwent surgical treatment and assigned them to the surgery group (S group). We then divided these subjects into those whose perforations were found to be covered intraoperatively (covered perforation group or S-C group) and those whose perforations were not covered (non-covered perforation group or S-NC group), based on their surgical records. Furthermore, the C and S-C groups were designated as the true conservative treatment group (new conservative group or N-C group), and the S-NC group was designated as the true surgical treatment group (new surgery group or N-S group). We then investigated a new scoring system using these groupings. We performed a univariate analysis using the Student t-test and Mann-Whitney U test to investigate the scoring items. We also performed multivariate analysis by means of logistic regression analysis for each item. We used a receiver operating characteristic (ROC) curve to calculate and investigate a cut-off value. A p value of <0.05 was considered statistically significant during each of the statistical analyses. We 56

3 perforated gastroduodenal ulcer 87 used JMP 12 (SAS Institute Japan, Tokyo, Japan) statistical analysis software. This study was approved by the ethics committee of the St. Marianna University School of Medicine (Approval No. 3376). II. Results The patient characteristics of the 117 patients with perforated gastroduodenal ulcers are shown in Table 1. There were 13 subjects (11.2%) in the C group, which comprised 8 men and 5 women with a mean age of 61.3 years. The etiology of the perforation was gastric ulcer in 6 and duodenal ulcer in 7 subjects. The outcome was survival in 12 subjects and death in one subject. Although conservative therapy was selected, one patient (7.7%) was converted to surgery due to worsening symptoms. The mean SIRS score was 1.0 ± 1.0, and the mean Watanabe score was 3.7 ± 0.9. The S group comprised 103 subjects (88.8%). Of these, 80 were men and 23 were women with a mean age of 63.4 years. The etiology of the perforation was gastric ulcer in 45 and duodenal ulcer in 58 subjects. A laparotomy was performed on 86 subjects (gastrectomy: 3 subjects, local surgery: 83 subjects), and laparoscopic surgery was performed on 17 subjects (all local surgery). The outcome was survival in 92 subjects and death in 11 subjects. There were no significant differences observed between the two groups in terms of sex, age, perforation site, or outcome. The mean SIRS score was 1.9 ± 0.9, and the mean Watanabe score was 1.8 ± 1.0. There was a statistically significant difference observed between the S and C groups in terms of the SIRS scores and the Watanabe scores (p < 0.05, Table 2). The 12 cases of mortality are summarized in Table 3 One subject in the C group died. The gastrointestinal perforation improved after conservative treatment, but the subject developed recurrent aspiration pneumonia due to dysphagia and died on day 64 of the hospital stay. The other 11 subjects died in the S group; all had a severe general condition during the initial examination such as cancer or severe sepsis. None of these conditions improved postoperatively, leading to the death of the subjects. The 103 subjects in the S group were divided into the S-NC group comprising 84 subjects (81.6%) and the S-C group comprising 19 subjects (18.4%). The details of each group are shown in Table 4. There were no statistically significant differences between the two groups in terms of sex, age, perforation site, and outcome. There was no statistically significant difference between the SIRS scores in the S-NC and S-C groups (2.0 ± 1.0 vs. 1.7 ± 0.7). However, there was a statistically significant difference between the Watanabe scores in the S-NC and S-C groups (1.6 ± 0.9 vs. 2.7 ± 1.2, p < 0.05, Table 4). The sensitivity and specificity of the Watanabe scores were 0.99 and 0.41, respectively (Fig. 2). From the 117 subjects with perforated gastroduodenal ulcers, we combined the 13 subjects in the C group and the 19 subjects in the S-C group to form the N-C group comprising 32 subjects and assigned the remaining 84 subjects in the S-NC group to the N-S group (Fig. 3). There was no significant difference between the SIRS scores in the N-C and N-S groups (1.5 ± 0.9 vs. 2.0 ± 1.0). However, there was a statistically significant difference between the Watanabe scores in the N-C group and N-S groups (3.2 ± 1.2 vs. 1.6 ± 0.9, p < 0.05) (Table 5). We then performed a multivariate analysis of each item in the Watanabe scoring system that showed a significant difference when we compared the N-S and N-C groups. We confirmed four items for which a significant difference was observed, and all four were risk factors, namely: time from onset to initial examination was within 6 hours (p = ), symptoms of peritoneal irritation were localized to the epigastrium (p = ), progression of ascites was limited to the epigastrium (p = ), and no Table 1. Characteristics of Patients with Upper Gastrointestinal Tract Perforation (n = 117) Sex ratio (male: female) 88: ± 16.3 (19-94) 52: 65 Outcome (survival: death) 105: 12 57

4 88 Niwa K Mikami S et al Table 2. Patient Characteristics Per Study Group Conservative Characteristic Surgery group therapy p (n = 103) group value (n = 13) Sex ratio (male: female) 80: 23 8: ± ± 18.3 (19-94) (36-90) : 58 6: SIRS Score 1.9 ± ± 1 (0-3) 0.02 Watanabe Score 1.8 ± ± 0.9 (0-5) (2-5) <0.001 Simple Open closure: 83 Operative procedure (n = 86) Gastrectomy: 3 Laparoscopy (n = 17) Simple closure: 17 Outcome (survival: death) 92: 11 12: Table 3. Patients Who Died During the Hospitalization Period and Pertinent Details Age (year s) Se x SIR S score Watana be Score Day of death Cause of death 43 M 3 2 POD 1 Ventricular tachycardia 69 M 2 3 POD 24 DIC 85 F 3 1 POD 28 DIC 88 F 2 0 POD 7 DIC 87 M 1 2 POD 8 Pulmonary embolism 63 M 0 1 POD 52 Sepsis 50 F 4 1 POD 20 Pulmonary failure 84 M 3 0 POD 22 DIC 76 F 3 1 POD 27 Internal bleeding 46 F 2 0 POD 14 Pulmonary failure 73 M 1 2 POD 7 DIC 88 M 0 3 Day 64 DIC: disseminated intravascular coagulation; POD: Dysphagia aspiration pneumonia postoperative day; SIRS: systemic inflammatory response syndrome. 58

5 Table 4. Characteristics of the Covered Perforation Group and Noncovered Perforation Group Characteristic Non-covered (n = 84) Covered perforation group perforation group p value (n = 19) Sex ratio (male: female) 64: 20 16: SIRS score Watanabe score perforated gastroduodenal ulcer ± 15.6 (19-94) 60.1 ± 19.2 (26-89) : 47 8: ± ± ± 0.7 (1-3) 2.7 ± 1.2 (1-5) Outcome (survival: death) 74: 10 18: SIRS: systemic inflammatory response syndrome. Fig. 2. The sensitivity and specificity of the Watanabe score SIRS: systemic inflammatory response syndrome. Fig. 3. The flow diagram showing allocation of patients with upper gastrointestinal tract perforation to therapy severe concomitant disease (p = ). No statistically significant difference was noted for the absence of a full stomach (p = ). When we established a coefficient based on the parameter estimated value for each factor that underwent logistic regression analysis, we found that the intercept was 0.36, and that 1.29 was to be added for symptoms of peritoneal irritation limited to the epigastrium, 0.92 to be added for the absence of severe concomitant disease, 0.80 to be added for time of onset to first examination of less than 6 hours, and 0.63 to be added for progression of ascites was limited to the epigastrium. In the ROC analysis (Fig. 4), the area under the curve (AUC) was 0.83, the cut-off value was 2.08, and the sensitivity and specificity were 0.94 and 0.63, 59

6 90 Niwa K Mikami S et al Table 5. Characteristics of the New Surgery Group and New Conservative Treatment Group Characteristic New surgery group (n = 84) New conservative treatment group (n = 32) p value Sex ratio (male: female) 64: 20 24: SIRS score Watanabe score 64.2 ± 15.6 (19-94) 60.6 ± 18.6 (26-90) : 47 14: ± ± ± 0.9 (0-3) 3.2 ± 1.2 (1-5) 0.1 < Outcome (survival: death) 74: 10 30: SIRS: systemic inflammatory response syndrome. Table 6. Characteristics of Patients Who Survived the Hospitalization Period and Those Who did not Characteristic Patients who survived patients who died p value (n = 105) (n = 12) Sex ratio (male: female) 81: 24 7: ± ± 17.0 (19-94) (43-88) : 59 6: SIRS score 1.8 ± ± Watanabe score 2.1 ± ± 1.1 (0-5) (0-3) 0.02 SIRS: systemic inflammatory response syndrome. respectively(fig. 6). III. Discussion Conservative treatment for perforated gastroduodenal ulcers was first reported by Wangensteen 8) in Thereafter, conservative treatment was gradually attempted more often, and a report was published by Taylor 9) in 1957 regarding 235 of 256 subjects who were treated conservatively. Due to the advent of H2 blockers and proton pump inhibitors, there have been remarkable changes in pharmaceutical therapy for peptic ulcers, and there were even scattered reports from Japan of conservative treatment for upper gastrointestinal perforations from the latter half of the 1908s 10). We also now know that Helicobacter pylori eradication therapy decreases the recurrence of ulcers. In conjunction with this, treatment for upper gastrointestinal perforation has shifted away from conventional gastrectomy; an increasing number of hospitals have been introducing minimally invasive treatments such as conservative therapy and laparoscopic surgery, and there have been numerous reports regarding their efficacy 11). The main option chosen in most of the reports is conservative treatment, but conversion to surgery occurs at a rate of 30%, which is considered to be high. If conservative treatment is successful, then this is beneficial, as the risks of surgery itself are thereby 60

7 perforated gastroduodenal ulcer 91 Fig. 4. ROC analysis of the new score The area under the curve (AUC) was 0.835, the cut-off value was 2.08, and the sensitivity and specificity were 0.94 and 0.63, respectively. Fig. 6. The sensitivity and specificity of the new score avoided. However, if there is conversion to surgery, then there is a high likelihood of the condition worsening from the time of examination. Accordingly, we cannot necessarily say that the high rate of selection of conservative treatment is a favorable finding while the rate of conversion to surgery remains high. We retrospectively investigated the outcomes of cases treated according to the criteria used at our hospital since When we investigated deaths, we found no cases of surgical death due to surgical manipulation or surgical invasiveness. In cases that were treated conservatively, the gastrointestinal perforation improved, but one subject died of recurrent aspiration pneumonia. Based on the Watanabe score, the rate of conversion to surgery was low at 7%. As a result, we believe that the Watanabe criteria that have been used to date can be said to have retained their validity. There was a significant difference observed between both the SIRS and Watanabe scores found in the C and S groups. However, when we compared surgical cases in terms of the presence of a covered perforation, we found no significant difference in the SIRS scores but did find a significant difference in the Watanabe scores. In the N-S and N-C groups, logistic regression analysis also showed no significant difference in terms of the SIRS scores, and only revealed a significant difference in the Watanabe scores. We believe that the SIRS score was not a useful way to divide the N-S and N-C groups. Rather, we believe that a new scoring system is needed that provides as easily and accurately as possible a method for differentiating the form of treatment required in patients who arrive at hospital with an acute abdomen. In the present study, if we exclude the SIRS score, which we determined was not a useful way to divide the two groups, and use only the Watanabe score, we predict that it will be effective to provide weighting for each item in such a way that enables a more definitive method of determining treatment. The ROC analysis to determine the cut-off values for the new scores resulted in a value of 2.08, and we have established that patients with a score of 2.08 or greater should be designated as those who should receive conservative treatment. When we repeated the determination of treatment in 116 subjects using the new scoring system, conservative treatment was selected for 25 subjects and of those, 5 subjects (20%) met the indications for surgery but had been selected to undergo conservative treatment. In addition, conservative treatment could have been selected for 12 of 91 subjects (13.2%) from the S group who were actually treated surgically. The current Watanabe score has a very high sensitivity of 0.99 and a specificity that does not exceed The sensitivity of the new score is approximately unchanged at 0.94, but the specificity is increased to This shows that the proportion of patients who are inadvertently treated by means of conservative treatment despite meeting the indications for surgery will not increase at our hospital when we use the current Watanabe score. It 61

8 92 Niwa K Mikami S et al is also possible to reduce the proportion of cases that are determined to meet the indications for surgery but that can actually be treated conservatively. In other words, we believe that the new scoring system used at this hospital at present is not only convenient, but it also enables a more accurate method of determining whether the patient meets the indications for surgery. In spite of this, we have theoretical concerns regarding the possibility of determining conservative treatment to be appropriate in patients who actually meet the indications for surgery. There are reports stating that there is an overproduction of cytokines during the clinical course of gastrointestinal perforations; this progresses to SIRS and the patient enters a state of shock 12). If the perforation occurs in the upper gastrointestinal tract, the peritonitis that occurs immediately after perforation is aseptic and is mainly a chemical peritonitis. The bacterial detection rate in ascitic cultures was reported to be 24% within the first 6 hours, 57% within 6-12 hours, and 100% from 12 hours onwards. Furthermore, from 24 hours onwards, Gram-negative bacilli were detected in all cases 13), and thus, it is only valid to select conservative treatment in cases presenting within the first 12 hours. Accordingly, in cases in which conservative treatment is selected to treat perforated gastroduodenal ulcers, we believe that it is possible to correctly identify cases that require surgical treatment by performing a re-evaluation after 12 hours. The new criteria are shown in Fig. 5. In terms of the surgical technique, the Guidelines for Endoscopic Surgery (2008) state that the treatment of perforated gastroduodenal ulcers tends to be converted to laparoscopic local surgery and recommend laparoscopic surgery, except in certain cases 14). We have also been proactively performing laparoscopic surgery to treat upper gastrointestinal perforations at our hospital since Even though this treatment was based on the Watanabe scoring system, and although it was possible to treat several patients with conservative treatment, there were also patients who required surgical treatment. Therefore, we believe that further investigation is required to determine the appropriate timing for abandoning conservative treatment. IV. Closing Remarks We conducted a retrospective study of treatment selection for gastroduodenal perforations determined by the Watanabe scoring system used at our hospital. Fig. 5. The flow diagram of the new scoring system Although the scoring system allows quick evaluation and is minimally invasive, there are many observation items, and it is somewhat complicated. We established a new scoring system based on the outcomes of this investigation, which simplified the previous scoring system, and we believe that this new system will be useful to further increase the precision of determining the indications for surgery. V. Acknowledgement The authors thank Professor Takahiko Ueno for his help in statistical analysis. We also thank Tina Tajima for English editing. References 1) Sunose Y, Ogawa T, Andoh T, Tomizawa N, Tanaka T, Sakamoto I, Motegi Y, Onozato Y, Owada S, Ikeya T, Takeyoshi I: A Clinical Study of Conservative Treatment for Perforated Gastric and Duodenal Ulcers [in Japanese with English abstract]. Journal of Abdominal Emergency Medicine 2006; 26: ) Omori H, Sasaki A, Ikeda K, Kawamura H, 62

9 perforated gastroduodenal ulcer 93 Koeda K, Kashiwaba Y, Wakabayashi G. Indication and limitation of non-operative management for perforated duodenal ulcer [in Japanese]. Journal of Clinical Surgery 2006; 61: ) Nagano M, Shimayama T, Takahashi N, Imamura N, Kawano K, Chijiiwa K: Indication of Conservative Treatment and It s Usefulness for Perforated Duodenal Ulcers[in Japanese with English abstract]. The Japanese Journal of Gastroenterological Surgery 2006; 39: ) Iwasaki K, Fukushima R, Inaba T, Morita N, Ikeda Y, Okinaga K, Takada T. Nonsurgical Treatment of Acute Perforated Duodenal Ulcer[in Japanese with English abstract]. Journal of Abdominal Emergency Medicine 2006; 26: ) Huke T, Okinaga K, Yokohata N, Fukushima R, Tomioka M, Wakakuri N, Shiraishi M, Yago T. Journal of Japanese College of Surgeons 1997; 22: ) Japanese Society of Gastroenterology Guideline Committee. Clinical Guidelines for Gastroenterological Tumors. Tokyo: Nankodo, ) Watanabe T, Shimada J, Katayama M, Enomoto T, Jinnouchi Y, Sakurai J, Hamaya M, Tobe N, Koizumi S, Asano T, Simamura T, Suda T, Asakura T, Shigeta H, Tanaka K, Tsukikawa S, Tanaka I, Otsubo T. New Criteria for the Clinical Treatment of Perforated Gastric or Duodenal Ulcer[in Japanese with English abstract]. Journal of Abdominal Emergency Medicine 2006; 26: ) Wangensteen OH: Non-operative treatment of localized perforations of the duodenum. Minn Med 1935; 18: ) Taylor H: Non-surgical treatment of peptic ulcer. Gastroenterology 1957; 33: ) Murata S, Kiyosaki K, Wakasa R, Ushijima S, Sugiyama S, Bandoh H. Surgery 1987; 49: ) Mouret P, François Y, Vignal J, Barth X, Lombard-Platet R: Laparoscopic treatment of perforated peptic ulcer. Br J Surg 1990; 77: ) Sugiyama M, Moriwaki Y. Perforated peptic ulcer. The Japanese Journal of Acute Medicine 1999; 23: ) Fujisaki M, Uematsu Y, Kurihara E, Kikuchi K, Tanimura C. Conservative therapy of perforated duodenal ulcer and gastric ulcer[in Japanese]. Surgical therapy 1988; 59: ) Japanese Society for Endoscopic Surgery. Clinical Guidelines for Endoscopic Surgery Tokyo: Kanehara Publishing Co; 2008;