Dr A. Mizuki, Department of Internal Medicine, Saiseikai Central Hospital, Tokyo, Japan. E-mail: email@example.com
Background : There are no management criteria for optimum out-patient care in mild-to-moderate acute colonic diverticulitis.
Aim : To enable such patients to be managed in an out-patient setting, by establishing criteria and treatment protocols.
Methods : We conducted an open trial and follow-up study from 1997 to 2002. On the basis of ultrasonography, we defined and categorized mild-to-moderate acute colonic diverticulitis ranging from limited inflammation within diverticulum to an abscess < 2 cm in diameter. Subjects were treated as out-patients and followed a 10-day treatment protocol consisting of an oral antibiotic and a sports drink for the first 3 days. Physical examination and laboratory testing helped determine whether or not a patient could resume a liquid diet on day 4, and a regular diet on day 7.
Results : Of the 70 patients, 68 were successfully treated. Two patients required hospitalization. Of the 65 patients who were tracked over several months [median (intraquarter range) = 30.8 (11.9–44.2) months], 16 had one or more clinical recurrences. The medical cost per episode was 80% lower than in-patient treatment.
Conclusions : Patients with mild-to-moderate acute colonic diverticulitis can be safely and successfully treated as out-patients using this protocol.
As the incidence of diverticular disease is increasing, and the prevalence increases with age,1 establishing a management strategy for acute colonic diverticulitis (ACD) is a priority. Most patients with ACD have a mild-to-moderately severe form of the disease, and few cases require surgical intervention.1 At present, there is a variety of management strategies that are used to treat such patients and these strategies include in-patient care with bowel rest and intravenous antibiotics.1 The need for hospitalization is usually determined on clinical grounds. There are no generally accepted criteria or classifications to guide the management choices when considering hospitalization because, in the past, the main focus of concern and study was deciding whether or not surgical intervention was indicated.2
It is generally accepted that the majority of cases with small confined pericolonic abscesses can be treated through conservative treatment methods.1 In this study, therefore, we defined cases with small confined pericolonic abscesses as borderline. Although computerised tomography (CT) is considered by some to be the most accurate and effective method of diagnosing ACD, particularly on the left side (and this modality can also be effective in detecting pericolonic abscess),3–8 recent studies suggest that ultrasonography (US) can give an accurate diagnosis.9–14 Based on this, and the assumption that these patients would be treated in clinics where US rather than CT was available for diagnosis, we chose US as the main diagnostic tool. In addition, US is more useful in monitoring the patient's response to treatment.13
On the basis of US findings, we differentiated between those patients with ACD who could be treated in an out-patient setting and those who required in-patient care. We conducted a trial of out-patient therapy using a standardized protocol consisting of an oral antibiotic and a sports drink. The patients were tracked and the recurrence rate was determined. On the basis of our findings, we have proposed a treatment scheme for patients with mild-to-moderate ACD.
This was a prospective study and an open trial. The study protocol was approved by the review board of Saiseikai Central Hospital, and all patients gave written informed consent.
All out-patients who were under 80 years old and had been diagnosed with ACD were eligible to be enrolled as study subjects. The clinical diagnosis was made on the basis of symptoms, physical examination, blood tests and blood culture results. In patients suspected of having ACD, the diagnosis was based on US findings and the degree of ACD was defined as either mild-to-moderate (grade I) or severe (grade II) (Figure 1). US was performed by expert technicians, and a diagnosis was made by three Radiologists. Following is the classification method we devised: grade I was classified into three categories, in which Ia indicated an inflamed diverticulum, where the inflammation was limited to the diverticulum; Ib indicated an inflamed diverticulum with pericolitis (inflammation clearly extended to the pericolic tissue around the diverticulum) and Ic indicated an inflamed diverticulum with an abscess of 2 cm in diameter or less. Grade II indicated an inflamed diverticulum with an abscess of >2 cm in diameter or with perforation into the abdominal cavity. When the US image resolution was poor (due to obesity or intestinal gas), a CT was also performed to confirm the diagnosis and the grade.
The following condition made patients ineligible to participate in this study: grade II ACD; those with bacteraemia and any severe comorbidity (such as uncontrolled diabetes mellitus, heart failure, renal failure or end-stage cancer). Also excluded was anyone who had undergone pre-treatment involving broad spectrum antibiotics (within 24 h prior to presentation in the out-patient clinic), and anyone who was considered to be unable to understand the protocol or manage self-care.
Treatment protocol and follow-up procedures
The flow chart for the out-patient treatment protocol and the follow-up procedures are summarized in Figure 2. The 10-day treatment protocol consisted of an oral antibiotic (cefpodoxime proxetil 200 mg twice daily for 10 days) with at least 1500 mL/day of a sports drink (405 kcal, 27 kcal/100 mL) for the first 3 days. A free intake of water was allowed during the first 10 days. If the patient showed improvement on day 4, a clear liquid diet was allowed. If improvement was still evident on day 7, a regular diet was resumed (patients were also advised to increase their fibre intake). Where there was no improvement, the patient was hospitalized and given intravenous antibiotics. The evaluation on day 4 and day 7 included a physical examination and blood tests, particularly a white blood cell (WBC) count and a C-reactive protein (CRP) level. A repeat US was performed on day 4. The decision on whether to continue the protocol on day 4 and day 7 was made on the basis of the patient's symptoms and a physical examination. After clinical signs of local inflammation had disappeared, a final evaluation was performed. This evaluation included a physical examination and a barium enema to confirm the presence of diverticula and rule out colon cancer.
Patients who had successfully completed the protocol participated in a follow-up study lasting until July 2002. The same criteria that had been used in the initial diagnosis of ACD were applied when diagnosing recurrences. A recurrence was treated using the same protocol. Where study participants had not been seen during the follow-up period, a telephone interview was conducted at the end of this period to identify whether they had experienced any recurrences.
Medical cost analysis
The direct cost of both out-patient and in-patient care for ACD patients was calculated from a review of the medical records. The direct medical cost of the out-patient care included doctor fees, diagnostic procedures, laboratory testing, out-patient follow-up and medication costs. The medical cost for the treatment of the hospitalized patients included the ward accommodation charges and the costs of the medical procedures and devices, as well as the out-patient costs.
All the statistical analyses were performed using spss 10.1 (SPSS Japan, Inc., Tokyo, Japan). Kruskal–Wallis sum rank tests, a one-way anova with Bonferroni's multiple comparison and a chi-square test were used to assess the significance of differences between the two grades of ACD severity. The continuous data was expressed as either mean (s.d.) or median [intraquarter range (IQR)]. In the follow-up study, the primary end-point was defined as a recurrence of ACD. In order to investigate the factors that affect this event and to calculate the accumulating event-free interval, time-to-event methods were used. Time-to-event was defined as the number of months (i.e. days/30) between the end of the treatment period and the day of the recurrence. The accumulated ACD-free intervals were calculated using the Kaplan–Meier method. Statistically significant differences in the event-free time between the groups, stratified by the affecting factors, were determined by a log-rank test.
Baseline characteristics of participants
Initially, 137 of the patients who were screened between July 1997 and April 2002 had ACD (Figure 3). The mean age of these 137 patients was 47.3 years. Eighty-four of these patients were male and 53 were female: 20 patients were diagnosed with grade Ia ACD, 101 with grade Ib, 10 with grade Ic and six with grade II. Of these 137 patients, 54 were excluded from the treatment protocol study: six because of grade II severity, nine because of other comorbidities, three because of pre-treatment involving antibiotics and 35 either because of perceived social problems or because they were unable to understand the protocol or to manage self-care. In one patient, the diagnosis was changed to appendicitis in subsequent assessment. Of the 83 patients that met our criteria, seven refused to give informed consent. There were no cases of patients exhibiting bacteraemia. The 76 patients who agreed to participate in this study entered the protocol. However, the data from six were excluded from subsequent analysis: four of the six had no visible diverticula on the barium enema examination; of the other two, the day 4 US showed that one had acute appendicitis and the other had colon cancer.
Of the 70 participants who had analysable data (Table 1 shows the baseline characteristics), the data was categorized according to ACD severity. The age range of the patients was 25–79 years [mean (s.d.) =47.8 (13.2)], with a peak age of 50. Thirteen had grade Ia ACD, 54 had grade Ib ACD and three had grade Ic ACD. In 61 patients, the ACD was on the ascending colon, in six on the descending colon and in three on the sigmoid colon. There were no patients with ACD of the transverse colon. All patients had abdominal pain, 22.9% of the patients had diarrhoea, 10.0% had experienced nausea, 2.9% had experienced vomiting and 11.4% complained of loss of appetite. There was a strong correlation between the severity of the ACD (based on US diagnosis) and the level of CRP, but not with the WBC or with other clinical manifestations.
Table 1. Baseline characteristics of patients by grade of colonic diverticulitis (CD)
Grade Ia (n = 13)
Grade Ib (n = 54)
Grade Ic (n = 3)
N.S., not significant; ND, not determined; A, ascending colon; D, descending colon; S, sigmoid colon; WBC, white blood cell; CRP, C-reactive protein.
a Median (intraquarter range).
b One-way anova.
c Chi-square test.
d Kruskal–Wallis test.
History of colonic diverticulitis
12 000 (7700–13 550)a
10 450 (9000–13 150)a
10 300 (8950–ND)a
Loss of appetite
The clinical outcomes are summarized in Figure 3. Of the 70 patients who were enrolled in the protocol, 68 patients completed the protocol without complications. A 59-year-old male patient with grade Ib ACD was removed from the study because of poor compliance with medication after being assessed on the day 4. A 48-year-old female patient with grade Ib ACD showed worsening symptoms (increased abdominal pain, fever and tenderness) during the out-patient treatment, so she was admitted on the day 7 of the protocol. Subsequent US showed that the patient had at least three inflamed diverticula on the ascending colon. However, all of the 68 patients who were successfully treated had less than three inflamed diverticula.
Recurrence of acute colonic diverticulitis
Complete follow-up information was obtained for 65 of the successfully treated patients. We were unable to obtain any information about the incidence of recurrence in three patients following the first treatment protocol. Each of these three patients had grade Ib ACD at presentation. The median follow-up period was 30.8 months (IQR = 11.9–44.2, range: 0.9–58.0). During the follow-up period, 16 patients had a recurrence of ACD symptoms. Of those 16, 15 had grade Ib severity in the initial attack (the one exception had grade Ic severity). Twelve of these 16 patients agreed to continue their participation in this study. In eight of these 12 patients, the recurrent ACD was in the same region as the first attack and four of these 12 patients had grade Ia severity while the others had grade Ib severity. All 12 patients repeated the treatment protocol and the ACD subsided. Two of these 12 patients had a third episode (grade Ib in severity), which was treated using the protocol.
The accumulated ACD-free time was 47.4 months (95% CI: 42.2–52.7) in patients with no history of ACD and 32.7 months (95% CI: 18.9–46.6) in patients with a past history of ACD. This difference did not reach statistical significance (P = 0.083) when assessed using the log-rank test.
Of the 54 ACD patients who were excluded from participation in this study, 45 who had been admitted to hospital gave consent for a medical cost analysis. These included one patient with grade Ia severity, 35 patients with grade Ib, four patients with grade Ic and five patients with grade II. All those with grade II did not improve with medical care (hospitalization with intravenous antibiotics), and required surgical intervention. The median (IQR) length of hospital stay was 9.0 days with grade Ia, 8.6 days with grade Ib, 7.2 days with grade Ic and 20.4 days with grade II. The median total cost of the treatment during hospitalization with either intravenous antibiotics or an operation (or both) was $1821 for grade Ia severity, $2068 for grade Ib, $2011 for grade Ic and $5571 for grade II (1$ = 110 yen). However, the cost of the protocol was $343 – <20% of the cost for conventional in-patient care of ACD patients.
This present study has demonstrated that US findings were successful in identifying those patients with a mild-to-moderate degree of ACD who can be treated in an out-patient clinic setting. Our study also showed that using the protocol in an out-patient clinic was safe and that there were no complications.
All of the patients with grade I severity responded to the protocol, whereas the patients with grade II ACD did not improve with medical care after being hospitalized, and required surgical treatment. The scheme devised by Hinchey et al.2 is useful in classifying the various inflammatory conditions in patients with ACD. However, the main focus of Hinchey et al.’s.2 classification is to identify patients with perforated diverticula, for whom surgical treatment should be recommended. It is not generally useful in assessing patients with mild-to-moderate ACD. With our US classification it is possible to differentiate between Hinchey's stage I and II2 and to judge the necessity of admission.
There is some debate as to the most appropriate diagnostic modalities to use in the accurate assessment of ACD, with both CT and US having some advantages.3 CT is considered to be the gold standard as a diagnostic procedure and this modality has additional potential in the detection of abscesses.3–8 However, several authors have recently suggested that US may also give a more accurate diagnosis and that it might be more useful in monitoring the response to treatment.9–14 Previous studies using US have suggested that this modality has a sensitivity of 84–98% and a specificity of 80–99%. There is a positive predictive value of 76–96% for US and 81% for CT; a negative predictive value of 84–96% for US and 81% for CT and an accuracy of abscess detection of 90–97% for US.3, 10–12, 14 These data suggest that US could be useful in the detection of smaller or earlier ACD lesions. In addition, US is inexpensive, non-invasive and is a widely available option.9, 12, 14 Although disadvantages of US include the poor image quality in obese patients, and in patients with excess colonic gas, poor image quality was only a minor problem in this study [seen in 2.9% (four of 137) of cases, so CT scan was performed] and the false-positive rate was 6.6% (five of 76). The false-negative rate in this study could not be assessed accurately, but there were no patients who needed any treatment after being diagnosed as negative for diverticulitis. We believe that these misdiagnosis rates are acceptable when using an initial diagnostic tool.
Previous investigators have recommended bowel rest for 3 days, with antibiotics and hydration for 7–10 days as an out-patient treatment for patients who are able to tolerate oral hydration.15, 16 Solid food was withheld for the first 3 days so that the bowel could rest. The ideal antibiotic would give broad spectrum antimicrobial cover, including anaerobes and Gram-negative rods.17 Our present protocol is consistent with these previous management strategies.
Recurrence is a serious problem in the management of ACD.18 In this study, 16 patients (23.5%) had a recurrence during the follow-up period. A number of previous reports have suggested that the recurrence rate after conservative treatment ranges from 20 to 60%.19–23 In comparison with these reports, the recurrence rate was not particularly high in this study. In addition, the severity of the recurrence was comparable with that at presentation and all recurrent cases responded to the same protocol. However, the accumulated recurrence rate in patients with a history of ACD was up to 50% within 2 years. Although this did not reach statistical significance, probably because of the small number of cases resulting in less statistical significance, the recurrence rate in patients with a history of ACD was higher than that in patients without such a history (30% within 4 years). Some authors have suggested that surgical intervention should be considered in patients with a recurrence.24–26 Indeed, two of our study subjects experienced three episodes of ACD within 6 months. The occurrence of multiple recurrences during a short period may be a relative indication for surgical intervention.
In our series of cases, the patients were all Asian and it has been reported that the location of ACD in Asians differs from that in westerners.1 In Asians, ACD is generally found on the right side; in Caucasians, ACD is more commonly located on the left side. In our current study on Japanese patients, 84% of these patients had right-sided ACD. Markham and Li27 have reported that ACD on the right colon tends to follow a more benign course than that on the left. On the contrary, it has also been reported that right-sided ACD takes a more severe clinical course.28 Although the incidence of ACD might vary based on ethnicity, the subsequent clinical course depends on the degree of severity of ACD at the time of the initial diagnosis.27, 28 We believe that it is the severity of ACD, and not ethnicity, which is the most important factor in deciding whether the protocol can be applied. Consequently, we believe that this protocol would be applicable to westerners.
In order to examine the effectiveness of the protocol in treating patients with mild-to-moderate acute colonic diverticulitis in an out-patient care setting, we decided to set up an open study design as an initial approach. Further comparison with a conventional study method may be needed. However, based on our present study, we have proposed a 10-day treatment protocol for ACD as described in Figure 4. Patients with grade I ACD, and with one or two inflamed diverticula, should be able to comply with the protocol. The treatment is effective in cases with or without a previous episode of ACD. It is reasonable to base the diagnosis of severity on the US findings. However, in cases with frequent recurrences (even if these are mild-to-moderate in severity), surgical intervention should be considered.
Authors would like to thank the residents and staff of the Internal Medicine, Surgery and Radiology departments of Saiseikai Central Hospital for their help in the care of our patients. In particular, authors would like to mention Hiromi Nishiya MD; Tatsuhiko Hayashi MD; Shinsuke Funakoshi MD; Kazuto Fukui MD; Takeharu Shigematsu MD; Takahide Kikuchi MD; Kentaro Watanabe MD. In addition, authors would like to thank Tsuneyuki Takashina MD and Kenichi Kodera MD, who and provided diagnostic ultrasound services; Yutaka Mishima, Mikiko Nozaki, Eriko Honda, Seishi Sasahara and Kyoko Kanemaru, who performed the ultrasound examinations; and Mitsuo Kitahara MD, who supervised the study.
Akira Mizuki and Hiroshi Nagata originally developed the study, performed the data collection and wrote the paper. Hiroshi Nagata provided overall direction for the study. Masayuki Tatemichi performed the statistical analysis and contributed to the writing of the paper. Satoshi Kaneda developed the ACD classification and was responsible for making the ultrasound diagnoses. Nobuhiro Tsukada was responsible for the medical care of the study patients. Hiromasa Ishii and Toshifumi Hibi supervised the study.