Mortality risk stratification in patients with aGvHD using allo-GRRR-OH score

Acute graft-versus-host disease (GvHD) is a serious complication of allogeneic hematopoietic stem cell transplantation (allo-HSCT) and a major cause of morbidity and non-relapse mortality. Accurate human leukocyte antigen (HLA) matching, using reduced-intensity conditioning regimens, and anti-infectious prophylaxis improved GvHD rates. However, with an estimated 20% of patients requiring intense and aggressive immunosuppressive treatment and admission to the intensive care unit (ICU), transplant-associated GvHD remains a challenge. Severe complications may be induced by a range of causes, including the consequences of GvHD disease, opportunistic infections due to immunosuppression, or toxicity of anti-GvHD therapy. Previous reports indicate poor prognosis for patients requiring ICU admission and those with pre-existing comorbidities. However, it remains unclear whether the need for life-sustaining treatment remains the same throughout the course of the disease or may be futile at some point in the GvHD trajectory.

To address those questions, Claire Pichereau and colleagues analyzed the per-intensive care GvHD trajectories and their impact on mortality in critically ill allo-HSCT recipients. The results were recently published in the journal Bone Marrow Transplantation.¹

Methods

• Data from all patients aged ≥ 15 years who were admitted to the ICU of the Saint Louis Hospital (Paris, France) after allo-HSCT between January 2005 and December 2014
• Various stem cell sources were used
• GvHD prophylaxis
  • Cyclosporine A (CsA) alone
  • CsA plus methotrexate
  • CsA plus mycophenolate mofetil
• First-line treatment for aGvHD
  • Methylprednisolone (>1 mg/kg/day)
• Second-line treatment for GvHD
  • Varied
• aGvHD trajectory was established based on time of the ICU admission and GVHD occurrence. Depending on the trajectory patients were then divided into groups
  • Group 1: patients with no occurrence of GvHD before or throughout the ICU stay
  • Group 2: patients with aGvHD controlled at the time of ICU admission
  • Group 3: patients with uncontrolled aGvHD (active, stable, or worsening aGvHD)
  • Group 4: patients with newly diagnosed, never treated aGvHD throughout the ICU stay
• Allo-GRRR-OH score was derived using the independent mortality predictors identified in the multivariant analysis of Day 90 mortality
  • Zero points were assigned if the predictor was absent
  • One point was assigned if the predictor was present in mild/moderate form
  • Two points were assigned if the predictor was present in severe form (mechanical ventilation, dialysis, or uncontrolled aGvHD)

Patient characteristics

• In total, 191 (15.8%) out of 1,212 of patients who underwent allo-HSCT required admission to ICU
• Median age was 42 (27–56) years and the majority of patients were male (68.4%)
• In total, 130 (68%) patients developed aGvHD
  • Median time between transplantation and the first sign of aGvHD was 21 days (11–39)
  • Majority of patients (87%) developed aGVHD before the life-threatening complications and ICU admission
• The most common underlying hematologic diseases were
  • Acute myeloid leukemia/acute lymphocytic leukemia/myelodysplastic syndromes (55.3%)
  • Non-Hodgkin lymphoma/Hodgkin lymphoma (18.4%)
  • Multiple myeloma (5.5%)
  • Other (20.8)

Results

• GvHD trajectory was assessed for 187 patients
• Outcomes, which were significantly different between groups, are presented in Table 1

Table 1. Outcomes by the GvHD trajectory group

aGvHD, acute graft-versus-host disease; ICU, intensive care unit *One patient lost to follow-up after discharge from ICU.
Characteristics	Group 1: no aGvHD (n = 57)	Group 2: controlled aGvHD (n = 82)	Group 3: uncontrolled aGvHD (n = 31)	Group 4: untreated aGvHD (n = 17)
Patients discharged from ICU, %	82.5*	76.8	87.1	94.1
Patients alive at Day 90, %	70.2	56.1	25.8	58.8
Overall survival, %	42.1	41.5	6.5	35.3

• Univariate analysis demonstrated that time between hospital and ICU admission, the severity of organ dysfunction, kidney and liver dysfunction, a requirement for life-supporting therapies, and diagnosis of aGvHD before ICU admission were significantly associated with Day 90 mortality
• Majority of factors associated with Day 90 mortality were confirmed by multivariate analysis (Table 2) 

Table 2. Multivariate analysis of Day 90 mortality

CI, confidence interval; GvHD, graft-versus-host disease; ICU, intensive care unit; OR, odds ratio; SOFA, sequential organ failure assessment
	OR (95% CI)	p value
ICU admission for respiratory failure	2.18 (1.09­–4.38)	0.02
ICU admission for liver failure	11 (1.2–100.4)	0.03
Acute kidney injury at ICU admission	1.87 (0.82–4.30)	0.13
SOFA score at ICU admission	1.21 (1.06–1.34)	0.004
GvHD trajectories Group 1 Group 2 Group 3 Group 4	Reference 1.47 (0.67–3.22) 6.64 (2.23–19.67) 0.73 (0.19–2.70)	0.34 0.0007 0.64

• The independent prognostic factors were used to derive the allo-GRRR-OH score, which was able to stratify patients by Day 90 mortality
  • Score 0–1 with low mortality risk (29%)
  • Score 2–3 with intermediate-risk (65%)
  • Score 4–5 with high risk (90%)

Conclusion

The results of this study highlight the poor outcomes of patients with uncontrolled aGVHD at the time of ICU admission and provide a way to stratify patients based on mortality risk. In patients with high allo-GRRR-OH score, doctors may need to decide at some point whether to continue aggressive life-supporting measures with limited benefit. Nevertheless, the authors advise that clinicians should avoid early end of life decisions in these patients before they have had a time-limited trial of at least 14 days.

In contrast, the authors suggest to promptly admit patients with intermediate-risk to the ICU and provide the best standard of care in order to improve mortality risk. However, in patients with low allo-GRRR-OH scores, the authors suggest that aggressive critical care management can be avoided. A prospective study is needed to verify the results in a bigger patient population.