Routine Evaluation of the Distal Colon Remnant Before Hartmann’s Reversal is Not Necessary in Asymptomatic Patients

Abstract

Background

Reversal of Hartmann’s is a common surgical procedure. Routine preoperative evaluation of the distal colonic/rectal remnant (DCRR) with contrast and/or endoscopic studies is frequently performed despite lack of evidence to support this practice. We hypothesize that asymptomatic patients can safely undergo Hartmann’s reversal without preoperative DCRR evaluation.

Methods

Adult patients undergoing reversal of Hartmann’s at a single institution were retrospectively identified. Operative characteristics and outcomes in patients with and without preoperative DCRR evaluation were compared.

Results

Between 1993 and 2008, 203 patients underwent reversal of Hartmann’s at a tertiary referral center. Sixty-eight patients (33%) did not undergo preoperative DCRR evaluation and had comparable demographic characteristics, comorbidities, DCRR length, and perioperative outcomes to 135 patients who underwent preoperative contrast and/or endoscopic studies. After evaluation, 125 (93%) patients had normal findings, seven (5%) patients had abnormal studies that did not impact their management, and three (2%) patients underwent additional procedures.

Conclusion

Hartmann’s reversal without previous DCRR evaluation is acceptable in selected asymptomatic patients, without increased risk of complications.

Introduction

Hartmann’s procedure involves segmental colonic resection with end-colostomy or end-ileostomy and closure of the distal colonic/rectal remnant (DCRR), which remains in the pelvis or abdomen as a blind-ending pouch (Hartmann’s pouch). This procedure is commonly performed in emergency situations in patients who require partial colectomy and are deemed to be at high risk of complications from a primary bowel anastomosis. The number of patients who undergo takedown of their stoma as a second-stage procedure varies between 56% and 100%.^1–5

Preoperative DCRR evaluation by means of contrast and/or endoscopic studies is routinely requested by many surgeons to exclude leak, stricture, inflammation, and tumors, which could preclude Hartmann’s reversal. DCRR evaluation is safe and has only minor disadvantages including cost, radiation exposure, and patient discomfort. However, there is no clear evidence that this practice affects surgical management or benefits patients. A previous study reported abnormalities in 16% of routine contrast DCRR studies, although these altered treatment in only a small minority of cases.⁶ In addition, the role of endoscopy in this setting has not been defined.

We undertook this study to evaluate our use of preoperative evaluation of the DCRR. We were specifically interested to determine if the findings of these routine studies changed our management of asymptomatic patients. We hypothesize that Hartmann’s reversal can be safely performed without routine preoperative DCRR evaluation in asymptomatic patients lacking other indications for preoperative investigation of the DCRR.

Materials and Methods

Patients undergoing Hartmann’s reversal at the University of Wisconsin were retrospectively identified after institutional review board approval. We were unable to retrospectively identify patients undergoing preoperative DCRR endoscopic and/or contrast studies due to lack of specific codes for these procedures. Hartmann’s procedure was defined as segmental colectomy, with or without resection of contiguous distal ileum, with proximal end-colostomy or end-ileostomy and closure of the DCRR. Defunctionalized distal colon and rectum extending proximal to the recto-sigmoid junction was defined as long DCRR. Patients younger than 18 years of age and those undergoing DCRR contrast or endoscopic studies more than 180 days prior to takedown were excluded from the study. Patients undergoing additional surgical procedures at the time of Hartmann’s reversal were analyzed separately.

Medical records were reviewed for patient demographic characteristics, indications for Hartmann’s procedure, operative characteristics, length of hospital stay, and procedure-related morbidity and mortality. Anesthesia records were retrieved to determine American Society of Anesthesiologists (ASA) score, estimated blood loss (EBL), and operative time. The ASA score was used to compare patient preoperative overall physical health.⁷ Operative time was defined as the time from skin incision to closure. When EBL was reported as “minimal,” a value of 50 ml was arbitrarily assigned to allow quantitative comparison between groups. Length of hospital stay (LOS) was defined as the time from operation to discharge from hospital. Return of bowel function was defined as the time to initiation of an oral diet consisting of unlimited liquids.

Radiology and endoscopy reports were reviewed to determine study indications, results, and type of contrast used. For contrast studies, a Foley catheter was inserted though the patient’s anus and contrast was infused to distend the DCRR for adequate visualization. The presence of pouch diverticulosis or polyps not requiring treatment before Hartmann’s reversal were defined as normal findings as they did not affect surgical management.

Complication grade was assigned using a scale ranging from 1 to 5 according to the system proposed by Mazeh et al.⁸ Briefly, this system grades inpatient and outpatient complications in the first 21 days after colorectal surgery. In order to allow comparison of patient groups, we recorded only the highest complication grade for each patient.

Student’s t test, Fisher’s exact test, and the Cochran–Armitage test were used to compare continuous variables, noncontinuous variables, and trends in ordinal variables, respectively, with p < 0.05 representing statistical significance. Means are represented ± standard deviation.

Results

Patient Characteristics

Between 1993 and 2008, 214 adult patients underwent reversal of Hartmann’s procedure at the University of Wisconsin. Eleven patients underwent pouch contrast or endoscopic studies more than 180 days prior to takedown and were excluded, yielding 203 patients for analysis. One hundred thirty-five patients (67%) underwent pouch contrast and/or endoscopic studies prior to takedown and 68 (33%) did not. There were no significant differences in age, gender distribution, time interval from Hartmann’s procedure to Hartmann’s reversal, DCRR length, or ASA score in patients with and without preoperative DCRR evaluation (Table 1). The proportion of patients undergoing takedown after Hartmann’s procedure for complications of diverticulitis was higher in patients with contrast/endoscopic studies of their DCRR but did not reach statistical significance (p = 0.052). Fourteen of 135 patients (10.4%) who underwent DCRR evaluation and 12 of 68 (17.6%) of those who did not, had a long DCRR (p = 0.18). There was no difference in the proportion of patients undergoing Hartmann’s reversal by a surgeon other than the one who performed their Hartmann’s procedure (p = 0.10).

Table 1.

Demographic Characteristics and Surgical Indications in Patients Undergoing Hartmann’s Reversal With and Without Preoperative DCRR Study (n = 203)

			Study		No study			p value
Age [mean years±SD (n)]
	All patients		57±16 (135)		55±16 (68)			0.41
	HR only			57±16 (98)		57±16 (38)			0.90
	HR+additional procedure			57±14 (37)		55±12 (30)			0.50
Gender [Male : Female]
	All patients		72:63 (135)		37:31 (68)			1
	HR only			56:42 (98)		21:17 (38)			0.85
	HR+additional procedure			16:21 (37)		16:14 (30)			0.47
Time to reversala [mean days±SD (n)]
	All patients		254±343 (126)		190±162 (65)			0.16
	HR only			237±233 (93)		159±91 (35)			0.06
	HR+additional procedure			303±549 (33)		227±213 (30)			0.48
Long DCRRb
	All patients		14/135 (10.4%)		12/68 (17.6%)			0.18
	HR			11/98 (11.2%)		6/38 (15.8%)			0.56
	HR+additional procedure			3/37 (8.1%)		6/30 (20.0%)			0.28
Different surgeonc
	All patients		67/135 (49.6%)		25/68 (36.8%)			0.10
	HR			48/98 (49.0%)		13/38 (34.2%)			0.13
	HR+additional procedure			19/37 (51.4%)		12/30 (40.0%)			0.46
ASA score (1–5)
	All patients
	n		129		68				0.32
		1		4		5
		2		74		28
		3		49		33
		4		2		2
	HR only
	n		94		38				0.84
		1		3		2
		2		54		19
		3		36		17
		4		1		0
	HR+ additional procedure
	n		35		30				0.23
		1		1		3
		2		20		9
		3		13		16
		4		1		2
Indications for Hartmann’s procedure
All patients				135		68
Diverticulitis				80 (59%)		30 (44%)			0.05
Trauma				9		7
Perforation				8		4
Ischemia				6		5
Iatrogenic injury				5		6
Obstruction				4		7
Anastomotic leak				6		4
Crohn’s				3		1
Clostridium difficile colitis				4		1
Hemorrhage				2		1
Unknown/other				8		2
HR only				98		38
Diverticulitis				53 (54%)		18 (47%)			0.57
Trauma				9		6
Perforation				7		2
Ischemia				6		0
Iatrogenic injury				5		3
Obstruction				4		4
Anastomotic leak				5		2
Crohn’s				2		1
Clostridium difficile colitis				3		1
Hemorrhage				2		0
Unknown/other				2		1
HR+ additional procedures				37		30
Diverticulitis				27 (73%)		12 (40%)			0.01
Trauma				0		1
Perforation				1		2
Ischemia				0		5
Iatrogenic injury				0		3
Obstruction				0		3
Anastomotic leak				1		2
Crohn’s				1		0
Clostridium difficile colitis				1		0
Hemorrhage				0		1
Unknown/other				6		1

SD standard deviation, HR Hartmann’s reversal, ASA American Society of Anesthesiologists, DCRR distal colon/rectal remnant

^aInterval between Hartmann’s procedure and Hartmann’s reversal

^bLong DCRR is defined as defunctionalized distal colon and rectum extending proximal to the rectosigmoid junction

^cAttending surgeon(s) performing Hartmann’s procedure and Hartmann’s reversal

Sixty-seven of 203 (33%) patients underwent additional surgical procedures at the time of Hartmann’s reversal (Table 2). Fifty-five percent of these patients (n = 37) had preoperative investigation of the DCRR prior to reversal while the remaining 45% (n = 30) did not. There were no significant differences in age, gender, interval of time from Hartmann’s procedure to reversal, DCRR length, or ASA score in those patients with and without preoperative DCRR evaluation (Table 1). Similarly, there were no differences in surgeon (p = 0.46) between these groups. We did find that the proportion of patients with a diagnosis of diverticulitis was significantly higher in patients with contrast/endoscopic studies of their DCRR (p = 0.01).

Table 2.

Additional Surgical Procedures at the Time of Hartmann’s Reversal in Patients Undergoing Hartmann’s Reversal With and Without Preoperative DCRR Study (n = 67)

	Study (n = 37)a	No study (n = 30)a
Hernia repair	12	7
Resection of proximal colon	8	3
Small bowel resection	3	1
Gynecological procedures	4	3
Cholecystectomy	1	2
Liver resection	0	3
Other procedures	16	16
Total	44	35

^aNumber of patients

Contrast/Endoscopic Studies

In 135 patients undergoing DCRR evaluation prior to takedown, 117 contrast studies and 22 endoscopic studies were performed (Table 3). Four patients (3%) underwent both procedures. No patients had symptoms suggestive of DCRR pathology. Specific indications for performing DCRR evaluation were identified in very few patients. Mean time to surgery was 39 days after contrast study and 48 days after endoscopy (p = 0.36). In all cases where DCRR evaluation was performed, this was as a routine preoperative study in asymptomatic patients, without any clinical or laboratory indications of DCRR abnormalities.

Table 3.

Contrast Used and Time from Study to Hartmann’s Reversal in Patients with Contrast and/or Endoscopic DCRR Studies

	N (%)		Days to reversal [mean±SD (n)]		p value
Contrast type	117		39±42 (114)		0.36
Barium		69 (59%)
Water-soluble		30 (26%)
Gastrograffin		21
Hypaque		3
Not specified		6
Unknown		18 (15%)
Endoscopic study	22		48±46 (22)
Contrast and endoscopic studies	4		N/A

DCRR distal colon/rectal remnant, SD standard deviation, N/A not applicable

Barium and water-soluble contrast were used for 69 (59%) and 30 (26%) contrast studies, respectively. The type of contrast used was not specified in 18 of 117 studies (15%). There were no complications related to pouch contrast or endoscopic studies. Overall, only 7% of all patients undergoing preoperative evaluation had findings on preoperative DCRR investigation. Clinical management was affected by the findings of DCRR investigation in 2% of all studies obtained (Table 4). Takedown was delayed in one case of pouch stricture and one case of active Crohn’s disease. One additional patient underwent stricture dilation prior to colostomy reversal.

Table 4.

Characteristics of Ten Patients with 13 Abnormal DCRR Studies

Patient	Age	Sex	Indication for Hartmann’s	Study type/result	Change in management	Complication grade/type
1	66	M	Diverticulitis	E/diversion proctitis	None	1/atelectasis
2a	78	M	Ischemic colitis	E/ischemia	None	N/A
				C/stricture	Dilation	1/atelectasis
3	79	M	Diverticulitis	E/diversion proctitis	None	2/wound bleeding
4a	26	F	Anastomotic leak	C/stricture	Reversal delayed	N/A
				E/diversion proctitis	None	2/pancreatitis
5	77	F	Diverticulitis	C/possible leak	None	2/delirium
6	28	M	Diverticulitis	C/possible stricture	None	1/wound dehiscence
7a	25	M	Crohn’s colitis	E/active Crohn’s disease	Reversal delayed	N/A
				E/diversion proctitis	None	3/wound abscess
8	68	M	Clostridium difficile colitis	C/presacral abscess	None	2/ileus
9	62	F	Diverticulitis	E/diversion proctitis	None	2/atelectasis
10	62	F	Ischemic colitis	E/diversion proctitis	None	4/hypotension

DCRR distal colonic/rectal remnant, M male, F female, E endoscopic study, C contrast study, N/A not applicable

^aPatient with more than one abnormal study

Perioperative Outcomes

One patient who self-discharged against medical advice on postoperative day 3 and one patient who underwent thoracoabdominal aneurysm repair during the same hospital stay were not included in the LOS and complication analyses. There were no significant differences in EBL, operative time, time to return of bowel function, complication grade, mortality, and LOS between patients with and without pouch studies prior to takedown (Table 5). This did not change when patients undergoing additional procedures were excluded from analysis. There was one death on postoperative day 21 as a result of an anastomotic leak in a patient with metastatic pituitary adenocarcinoma.

Table 5.

Outcomes in Patients Undergoing Hartmann’s Reversal

			Study [mean±SD (n)]		No study [mean±SD (n)]			p value
Estimated blood loss (ml)
	All patients		238±228 (134)		272±269 (68)			0.35
	HR only			223±229 (97)		215±197 (38)			0.86
	HR + additional procedures			278±223 (37)		343±328 (30)			0.34
Operative time (min)
	All patients		183±70 (124)		203±98 (65)			0.10
	HR only			170±62 (91)		182±74 (36)			0.38
	HR + additional procedures			219±78 (33)		231±117 (29)			0.65
Time to liquid diet (days)
	All patients		5.2±3.5 (129)		5.1±2.6 (66)			0.88
	HR only			5.2±3.7 (95)		5.6±2.9 (38)			0.68
	HR + additional procedures			4.9±2.9 (34)		4.4±2.0 (28)			0.40
LOS (days)
	All patients		8.4±4.6 (132)		8.2±5.0 (67)			0.83
	HR only			8.1±4.4 (96)		8.5±6.3 (37)			0.73
	HR + additional procedures			9.1±5.1 (36)		8.0±3.0 (30)			0.29
30-day mortality
	All patients		1 (0.7%)		0			-
	HR only			0		0			-
	HR + additional procedures			1 (2.7%)		0			-
Complications
	All patients
	None		62/135 (45.9%)		33/68 (48.5%)				0.67
	Gradea
		1		16		2
		2		36		22
		3		17		6
		4		3		5
		5		1		0
	HR only
	None		48/98 (49.0%)		18/38 (47.4%)				0.34
	Gradea
		1		13		1
		2		23		13
		3		11		2
		4		3		4
		5		0		0
	HR + additional procedures
	None		14/37 (37.8%)		15/30 (50.0%)				0.53
	Gradea
		1		3		1
		2		13		9
		3		6		4
		4		0		1
		5		1		0

DCRR distal colon/rectal remnant, SD standard deviation, LOS length of hospital stay, HR Hartmann’s reversal

^aAs defined by Mazeh et al.,⁸ only the highest complication grade recorded

Three patients who did not have contrast or endoscopic studies of their DCRR were found to have previously unrecognized intra-abdominal abscesses at the time of takedown. All three underwent abscess drainage and washout with reanastomosis in the same setting. The complication rate of ten patients with abnormal DCRR studies was 100%, significantly higher compared to patients with normal DCRR results (50%; p = 0.002). With the exception of one patient who underwent a negative exploratory laparotomy for persistent unexplained hypotension, these were all grade 1–3 complications.

Discussion

Reversal of Hartmann’s is a common surgical procedure with significant morbidity.⁹ Previous reports have identified a number of abnormalities that can affect the DCRR and could potentially cause significant complications after Hartmann’s reversal, such as leak, fistula, stricture, malignancy, and diversion colitis.^3,6,10,11 These studies have generally been small retrospective series that, in many cases, have included patients with symptomatic conditions of their DCRR and/or those who were not candidates for reversal.^3,6,10,11 The largest report of asymptomatic patients revealed a low incidence of radiographic DCRR abnormalities.⁶ Therefore, in the absence of symptoms, the role of DCRR contrast and/or endoscopic studies before Hartmann’s reversal is unclear. In this study, we found that the postoperative outcomes of 68 asymptomatic patients undergoing Hartmann’s reversal without preoperative DCRR evaluation were equivalent to those of 135 patients cleared for surgery after contrast and/or endoscopic studies. This shows that DCRR evaluation prior to Hartmann’s reversal can safely be omitted in asymptomatic patients.

One of the controversies of using rectal contrast and endoscopy to evaluate the DCRR before Hartmann’s reversal is that their accuracy in this setting has not been studied. Although the present series did not specifically assess the accuracy of either type of study, no evidence of a leak was found intraoperatively in those patients with contrast studies suspicious for a leaking DCRR. Cherukuri et al. have also reported false-positive results with rectal contrast studies in the same setting. In another study, Da Silva et al. reviewed 84 preoperative contrast studies in patients scheduled for closure of diverting ileostomy after colonic J-pouch anal anastomoses and reported four false-positive results (three strictures and one leak) in a total of six abnormal studies.¹² Despite the differences in anatomy between a J-pouch anal anastomosis and a DCRR, this is additional evidence that rectal contrast studies are not accurate in identifying leaks and strictures. In addition to the risk of false-positive results, true-positive findings can sometimes negatively impact patients. Diversion proctocolitis, an inflammatory condition that affects defunctionalized bowel, can be misdiagnosed as inflammatory bowel disease and unnecessarily delay colostomy takedown.¹³ Diversion proctocolitis commonly affects the DCRR, but it is not an indication for preoperative endoscopy nor should it affect the management of asymptomatic patients who are candidates for Hartmann’s reversal.^10,13

Furthermore, the clinical significance of abnormal DCRR radiographic or endoscopic findings and their value in patient management is unclear. In 135 patients undergoing Hartmann’s reversal, preoperative DCRR evaluation by means of contrast and/or endoscopic studies revealed abnormalities in ten (7%) patients and altered the management of only three of them (Table 4). Our findings are corroborated by the report of Cherukuri et al. who found that, in 70 asymptomatic patients with 11 abnormal contrast studies of their DCRR, patient management was changed in three cases.⁶ Similar to our study, all three patients subsequently underwent successful Hartmann’s reversal.⁶

In the absence of findings on preoperative DCRR investigation that impact management, a DCRR study may still be indicated if preoperative planning leads to decreased complications. In fact, we found no difference in complication rate between those patients with and without DCRR investigation. Interestingly, all patients with abnormal DCRR studies had postoperative complications. Nevertheless, all these complications were minor and, given the small number of patients with abnormal studies, we do not believe that a cause-and-effect relationship between an abnormal DCRR study and postoperative complications can be inferred. Therefore, we do not believe that DCRR investigation is warranted for either decreasing complications or for changing management in the asymptomatic patient.

However, we do believe that DCRR evaluation prior to Hartmann’s reversal has a role in the management of some patients. Certainly, any patient with symptoms or signs of DCRR abnormalities should have DCRR investigation prior to reversal. In a previous report, five of 14 patients with symptoms indicating abnormalities of the DCRR were found to have abnormal DCRR contrast studies, which in all five cases affected patient management.⁶ In 24 patients with bleeding and/or pain affecting the DCRR, Haas et al. diagnosed 11 cases of severe proctitis, two polyps, and eight carcinomas, which were not further described in their series.¹⁰

We also believe that endoscopy of the remaining colon and rectum should be offered preoperatively to all patients who meet screening criteria for colorectal cancer. In their series of 25 asymptomatic patients who underwent DCRR evaluation at 1 year or longer after Hartmann’s procedure, Haas et al. diagnosed two polyps and one carcinoma.¹⁰ Due to the retrospective nature of this study and the changes in colonoscopic screening criteria, we were unable to determine the number of patients in our study population who met the screening criteria at the time of their operation.

A third group of patients who may have indications for routine preoperative DCRR evaluation is those patients with a diagnosis which carries a high risk of DCRR abnormalities. For example, a DCRR study in the Crohn’s patient is warranted to ensure disease in the distal remnant is controlled. In our study, one patient of three with Crohn’s disease was found to have acute inflammation requiring medical therapy. Similarly, a previous leak from a colonic anastomosis requiring Hartmann’s procedure may increase risk of DCRR stenosis and evaluation before takedown may be warranted. In this study, one of six patients with a history of anastomotic leak was found to have a DCRR stricture at the time of DCRR evaluation.

Finally, DCRR evaluation may be indicated in some cases to assist operative planning by defining its anatomy, especially in patients who had their Hartmann’s procedure by a different surgeon. Surgeons who are not comfortable performing low pelvic anastomoses would be more likely to refer patients with short DCRR to a surgeon with more experience. In our series, 45% of patients had Hartmann’s reversal by a different surgeon than the one who performed their Hartmann’s procedure. These patients were not more likely to have a preoperative DCRR contrast or endoscopic study.

A potential limitation of this retrospective series is that it may have not captured patients with abnormal DCRR studies who subsequently did not undergo reversal due to an abnormal study result or other reasons. However, patients who do not undergo reversal after Hartmann’s procedure are usually unfit for surgery or have a poor prognosis due to known malignancy and are unlikely to undergo DCRR endoscopic or contrast studies in the absence of symptoms.¹⁴ Furthermore, abnormalities of the DCRR do not preclude Hartmann’s reversal. In the study of Cherukuri et al., none of the 16 patients with abnormal DCRR contrast studies were turned down for Hartmann’s reversal as a result of their contrast study results.⁶ Similarly, we found that all patients with abnormalities on preoperative DCRR investigation were ultimately reversed. Furthermore, the retrospective nature of our study did not permit accurate identification of the indications for DCRR evaluation. Due to the lack of guidelines for asymptomatic patients, we believe this is largely determined by personal preference. DCRR length, time since Hartmann’s procedure, and previous Hartmann’s procedure by a different surgeon greatly influence the decision to study the DCRR preoperatively. Finally, lack of differences in patient outcomes could be related to small sample size, particularly in subgroup analyses.

Conclusion

Hartmann’s reversal without previous DCRR evaluation is acceptable in selected asymptomatic patients, without increased risk of complications. Contrast and/or endoscopic studies should be reserved for patients with symptoms indicating potential DCRR pathology, those patients who have situations predisposing them to complications of the DCRR, and those who meet the screening criteria for specific colorectal pathology.