Diverting enterostomy to treat steroid-resistant gastrointestinal acute graft-versus-host disease

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative therapy for a variety of hematologic malignancies; however, up to 50% of recipients develop acute graft-versus-host disease (aGvHD). aGvHD involving the gastrointestinal (GI) tract (GI-aGvHD) is difficult to treat, and in its most severe forms it carries a mortality rate of >95%. Any part of the GI tract may be affected, but the ileum is most commonly involved due to its high density of lymphoid follicles.

Surgical intervention in patients with GI-aGvHD is usually restricted to emergency presentations such as bowel perforation, GI hemorrhage, and intestinal necrosis. However, Justine Khodr et al., in Annals of Surgery,¹ hypothesize that a proximal diverting enterostomy (DE), a procedure in which the bowel is brought out to the abdominal surface to form a stoma,² could improve management of severe steroid-refractory GI-aGvHD in the distal excluded ileocolonic segment. The procedure diverts fecal matter away from the affected bowel, reducing bacterial translocation (the passage of bacteria through gut mucosa to normally sterile tissues³), which is implicated in GI-aGvHD pathophysiology. The results of their retrospective study are summarized in this article.

Study design

• Patients with steroid-refractory Stage III or IV GI-aGvHD who had undergone allo-HSCT between August 2004 and December 2018 were identified, as shown in Figure 1.
• Imaging of the GI tract (upper and lower digestive endoscopy, video-capsule endoscopy and/or computed tomography (CT), and barium meal follow-through) and plasma citrulline were used to assess the extent and severity of intestinal damage.
• DE was not offered to patients considered too frail for surgery.

allo-HSCT, allogeneic hematopoietic stem cell transplant; DE, diverting enterostomy; GI-aGvHD, acute graft-versus-host disease involving the gastrointestinal tract; MDT, multidisciplinary team.
*Adapted from Khodr, et al.¹

Baseline characteristics

Patient characteristics, presented in Table 1, did not differ significantly between the medical group and the DE group, except for cytomegalovirus status, which was significantly higher in the DE group (p = 0.003).

Table 1. Patient characteristics and transplantation modalities*

ALL, acute lymphoblastic leukemia; AML, acute myelogenous leukemia; CLL, chronic lymphocytic leukemia; CML, chronic myelogenous leukemia; CMV, cytomegalovirus; DE, diverting enterostomy; HL, Hodgkin lymphoma; IQR, interquartile range, MDS, myelodysplastic syndrome; MM, multiple myeloma; MPD, myeloproliferative disorder; NHL, non-Hodgkin lymphoma. *Adapted from Khodr, et al.1 †Assessed 30 days before the transplant.
Characteristic, % (unless otherwise stated)	Medical group (n = 38)	DE group (n = 13)
Median age, years (IQR)	54.5 (41.0–60.0)	44.0 (38.0–49.0)
Female	36.8	53.8
Hematologic malignancy
MM	13.2	23.1
ALL	2.6	30.8
HL	0.0	7.7
NHL	15.8	0.0
AML	31.6	15.4
MDS	0.0	7.7
CLL	10.5	0.0
MPD	15.8	15.4
CML	10.5	0.0
Prophylactic digestive decontamination	44.1	69.2
Positive recipient CMV serology	31.6	69.2
Median plasma citrulline, mmmol/L (IQR)†	26 (19–32)	27 (17–33)

Characteristics of aGvHD, presented in Table 2, were comparable between both groups, with the exception of a significantly higher rate of proximal GI-aGvHD involving the jejunum in the medical treatment group (p < 0.0001).

Table 2. Characteristics of aGvHD*

aGvHD, acute graft-versus-host disease; DE, diverting enterostomy; GI-aGvHD, aGvHD involving the gastrointestinal tract; ICU, intensive care unit; IQR, interquartile range. *Adapted from Khodr, et al.1 †p < 0.0001.
Characteristic, % (unless otherwise stated)	Medical group (n = 38)	DE group (n = 13)
GI-aGvHD stage
Stage 3	28.9	23.1
Stage 4	71.1	76.9
Proximal GI-aGvHD (involving jejunum)†	74.9	7.7
Liver aGvHD stage
No damage	65.8	69.2
Stage 1	0.0	7.7
Stage 2	2.6	0.0
Stage 3	10.5	7.7
Stage 4	21.1	15.4
Skin aGvHD stage
No damage	34.2	46.2
Stage 1	18.4	7.7
Stage 2	18.4	30.8
Stage 3	18.4	7.7
Stage 4	10.5	7.7
No. of immunosuppressor lines
n = 1	8.8	0.0
n = 2	38.2	61.5
n = 3	35.3	30.8
n = 4	17.6	0.0
n = 5	0.0	7.7
Missing data	10.5	0.0
ICU stay during hospitalization	65.6	76.9
Median delay from transplantation to aGvHD, days (IQR)	30 (25–52)	30 (27–40)
Median plasma citrulline during aGvHD, mmmol/L (IQR)	4 (3–6)	3 (3–5)

Results

• Overall survival was significantly higher in the DE group, as shown in Table 3.

• DE reduced the risk of death (hazard ratio, 0.226; 95% confidence interval, 0.092–0.554; p = 0.0011).
• Lower infection-related cause of death was seen in the DE group (30.8%) vs the medical group (57.9%), although this did not reach statistical significance (p = 0.091).
• Both groups showed a dramatic decrease in citrulline levels when measured at the worst clinical intensity of GI-aGvHD.

Table 3. Overall survival and causes of death*

DE, diverting enterostomy; GI, gastrointestinal; OS, overall survival. *Adapted from Khodr, et al.1
Outcome, %	Medical group (n = 38)	DE group (n = 13)
1-year OS	5	54
2-year OS	2.5	31
Cause of death
Relapse	2.9	10
Hemorrhage (non-GI)	10.5	0
Pneumonia	22.9	20
Encephalitis	2.8	0
Staphylococcal sepsis	2.8	0
Liver failure	5.7	20
Heart failure	5.7	10
Respiratory failure	5.7	20
Kidney failure	5.7	10
GI complications	45.7	30.8
Obstruction	5.7	0
GI sepsis	34.3	20
GI hemorrhage	5.7	20
Overall infections	57.9	30.8

Severe postoperative complications occurred in three patients in the DE group, with one patient having volvulus of enterostomy requiring endoscopic treatment, the second developing cellulitis with enterocutaneous fistula leading to death 25 days post-operation, and the third experiencing GI perforation leading to death 35 days post-operation.

In the medical group, ten patients were considered for surgery but received medical treatment only: three were considered too frail for surgery, one improved with medical treatment, and six patients should have undergone DE but did not and all six died within the first year after GI-aGvHD onset.

Conclusion

DE significantly improved survival outcomes in patients with steroid-refractory GI-aGvHD compared to medical management alone and this supports previous findings from a smaller study.¹ The effectiveness of DE is further validated given the significantly higher rate of cytomegalovirus infection in the DE group, a well-recognized cause of mortality in GI-aGvHD. However, notable limitations of this study include the small sample size, retrospective nature, and the fact that DE was preferentially performed in patients with distal GI-aGvHD, meaning the two groups were not entirely comparable. Given the poor survival rate in the medical group combined with the dramatic drop in citrulline levels in both groups, indicating deep intestinal failure and risk of bacterial translocation, Khodr and colleagues propose that DE should be considered in steroid-resistant GI-aGvHD regardless of GI-aGvHD location; research to assess this is ongoing.