Graft versus host disease

Graft vs host disease (GVHD) is usually a complication of haematopoietic stem cell transplantation (HSCT) but can also result from blood transfusion, granulocyte transfusion, CAR-T cell therapy and rarely from solid organ transplantation. This beast is well known to trainees that practice in bone marrow transplant centres and their satellites, as there is often a community of chronic sufferers around, and a large haematology/oncology inpatient population that circulates through the ICU.

It would be less known to those trainees that have little access to bone marrow transplant units, or whose experience of BMT complications might have been patchy. For the record, the Australian Bone Marrow Donor Registry lists only ten major hospitals around Australia as adult bone marrow transplant centres; whereas at the time of writing CICM lists 60 ICUs as accredited for "core" ICU training. It is therefore hardly surprising that only 25% of the trainees passed Question 25 from the second paper of 2022, which asked for detailed information about graft versus host disease.

It would of course, be unsurprising for niche esoteric topics to appear in this exam, even though they might only be seen in a few centres (hello, APRV). The argument for including them is surprisingly solid. If we limited the CICM fellowship exam paper to only the sort of bread-and-butter stuff one might see in a small six-bed regional unit, then how can we (and society at large) be confident that we are producing fellows who are critical care experts, trained to keep their head about them when confronted with mindblowing weirdness, an elite team ready for anything. 

On the other hand, this extremely niche esoteric problem is unlikely to ever be managed by ICU specialists in isolation, in the sense that it would be extremely unexpected for an intensivist to take on the whole role of workup and management decisionmaking for patients suspected of having GVHD. A haematologist is always waiting by the phone somewhere, ready to give advice. In the same vein, it would be unexpected for an amateur intensivist blogger to ever produce any material about GVHD that would compare to professional literature on the subject, or rival the advice of a haematologist. It would therefore make sense for this resource to focus on the needs of the exam candidates, especially those who have no access to a bone marrow transplant unit, by summarising some kind of exam-answer-level concentrate of the topic, and finishing with some free resources if more depth is ever required for some reason. Choi et al (2010) or Yeh & Deeg (2021) are probably some of the better papers to be skimming through for this topic, and one notes with surprise that the UpToDate chapter appears to be available to the destitute public.

What kind of graft would attack the host?

Most people, when they think of organ transplantation, would think of solid organs (and indeed it appears that many of the candidates answering Question 25 were actually answering a question about solid organ rejection). However those are scenarios where the graft is benign; in the sense that a kidney is unlikely to attack the recipient (well, it could passive-aggressively stop filtering, but that is a whole separate issue). The main reason for this is that the kidney, heart or liver lack any mechanism to recognise and injure the host. Not so with the donated immune system elements that are introduced into the host in the context of a haematopoietic stem cell transplant. In this scenario, the graft has weapons, and can attack, and this is in fact the most common complication of bone marrow transplantation.

In addition to HSCT, graft versus host disease is also occasionally seen in the context of any other way T-lymphocytes can sneak into a person, such as:

  • non-leukodepleted blood transfusion
  • granulocyte transfusion
  • rarely, solid organ transplant (classically liver)
  • CAR-T cell therapy (though one must admit that in this case being attacked by the graft is sort of the point)

Pathophysiology of graft versus host disease

As it is not the objective of this chapter to turn the reader into an immunopathologist, the following is a shameless oversimplification designed to pass the CICM Second Part by bamboozling the examiners with intelligent-sounding terminology. The CICM trainee can virtually guarantee that their examiner will not have advanced knowledge of the subject because it is extremely unlikely that their preparation for the marking of the written or viva would have included extensive reading of the published literature, let alone professional expertise in this subject matter (as the Australia training programs for haematology and immunology are even more fearsome than the CICM program, and so it is unlikely that anyone would ever complete a dual CICM/haematology fellowship). It is therefore safe to limit one's view of GVHD pathophysiology to the following blurry image:

  • Phase 1: activation of antigen-presenting cells
    • The host experiences proinflammatory stimuli, such as the haematological malignancy itself, the tissue damage resulting from conditioning chemotherapy, and any intercurrent sepsis (thus intense conditioning regimens are a risk factor for GVHD)
    • This results in inflammatory cytokine release (TNF- α and IL-1, etc)
    • The host's own antigen-presenting cells are activated by these cytokines
  • Phase 2: activation of donor T cells
    • Donor lymphocytes (allogeneic donor cells) are infused into the host 
    • The antigen-presenting cells then interact with the donor T-cells, and donor T cells recognize alloantigens presented by these cells. Because these antigens do not resemble any molecules that were found in the donor, the T cells recognise these antigens as something foreign.
    • Following antigen recognition T cells express numerous costimulatory molecules that produce proliferation, differentiation and migration of T cells
  • Phase 3: cytotoxic and inflammatory effects
    • Cytotoxic donor T cells damage organs directly by lysing cells via Fas/Fas ligand and perforin/granzyme pathways
    • The inflammatory cytokines secreted by activated T cells recruit other effector cells to the target organs, and these cells cause further tissue damage
    • These cytokines can also activate antigen-presenting cells and enhance allopresentation

Question 25 from the second paper of 2022 only allocated 20% of the mark to this aspect, which suggests that the trainee is encouraged to know less detail, not more; but in case more detail is what they want, Reddy et al (2003) contains an excellent overview which does not overwhelm with immunology jargon. 

Risk factors for developing acute graft versus host disease

In a Harvey lecture from 1966, Ruper Billingham laid out the three main things that are required for a graft versus host reaction:

  • The graft must contain immunologically competent cells
  • The recipient must express tissue antigens that are not present in the transplant donor
  • The recipient must be incapable of mounting an effective response to eliminate the transplanted cells

Those are probably some kind of minimum, in terms of what might be considered risk factors, but probably a little more practical detail is required for answering exam questions. The following list of risk factors is from Przepiorka et al (1999), but with blended wisdom from other sources.

  • Unrelated donors
  • HLA mismatching
  • Donor/recipient sex mismatch (female to male is greatest risk)
  • Racial mismatch
  • Older age (over 40)
  • Aggressive conditioning
  • Total body irradiation
  • High CD34+ cell dose
  • Blood cell grafts (as opposed to bone marrow)
  • CMV positive status
  • Functional status (Karnofsky performance score less than 90, which is scored out of 100 where 100 is "totally normal" and 90 is merely "able to carry on normal activity; minor signs or symptoms of disease")

In general it appears to be extremely common and the risk for acute GVHD ranges from 32-38% in the best circumstances to over 60% in poorly matched donors.

Acute,  chronic and late-onset GVHD

The distinction of what "chronic" means in the fast-moving world of haem oncology is conventionally held to be 100 days post transplant, i.e. the GVHD that appears at 100 days was  called "chronic", whereas the same GVHD on day 99 is acute. More modern definitions focus on clinical features which are somewhat distinct between these entities. The 2005 NIH consensus criteria describe them as follows:

  • Acute GVHD: occurs within 100 days
  • Late onset GVHD: looks clinically like acute, but occurs after 100 days
  • Chronic GVHD: looks clinically distinct (nothing like acute GVHD), occurring at any time after transplant (not necessarily 100 days later)
  • Overlap syndrome: a mixture of acute and chronic features

So, the 100 day barrier is not really a barrier between the acute and chronic forms. ​​​​​​​Yeh & Deeg (2021) have an excellent breakdown of the types and timing in the form of a diagram, which is paraphrased here in a way which should hopefully help it become memorable and reproduceable for stressed exam candidates:

Classification of GVHD

There appears to be considerable overlap between these processes, in the sense that acute and chronic can merge into each other, but there appears to be something pathophysiologically distinct about them: Toubai et al (2008) describe chronic GVHD as more of an autoimmune condition, whereas acute GVHD is more like a transfusion reaction. This, as well as the chronicity of inflammation, produces a different spectrum of disease in chronic GVHD patients, which brings us to:

Clinical manifestations of GVHD

​​​​​​​Yeh & Deeg (2021) also have an excellent breakdown of the different clinical features detailed in the NIH 2014 report.  To summarise:

  • Acute GVHD:
    • Skin: inflammatory erythematous mculopapular rash
    • Gastrointestinal: nausea, vomiting, diarrhoea, ​​​​​​​anorexia, mucositis
    • Liver: elevated bilirubin
  • Chronic GVHD:
    • ​​​​​​​Absence of acute features (skin, gastrointestinal or liver)
    • Sclerotic skin disease with hyperpigmentation and ichthyosis
    • Lichenoid oral and genital mucosa
    • Gingival disease, oral pseudomembranes
    • Oesophageal strictures
    • Nail changes (dystrophy, ridging, etc)
    • Alopecia
    • Ocular changes (chronic conjunctivitis)
    • Myosits 
    • Bronchiolitis obliterans
    • Myositis and neuromyopathy
    • Pleural and pericardial effusions

Diagnostic evaluation of acute GVHD

"The rationale for diagnostic evaluation was commonly omitted", complained the examiners in their Question 25 from the second paper of 2022. It is not usual for one to have to justify the need to reach a diagnosis, as the desirability of a diagnosis is often self evident, but let's have a go anyway:

  • ​​​​​​​GVHD can be diagnosed clinically but the characteristic features are common manifestations of other illnesses prevalent in the ICU (eg. rash and diarrhoea due to antibiotics, hyperbilirubinemia due to cholestasis)
  • The management of GVHD is cardinally different to the management of many of its differentials
  • GVHD can become life-threatening if not identified and treated

Still, the diagnostic process needs to include clinical findings:

  • Histological confirmation is the gold standard (the affected organ is biopsied, to discriminate between organ-specific differentials, eg. where CMV colitis is an equally likely explanation for the diarrhoea)
  • Clinical criteria (especially when it comes to the diagnosis of chronic GVHD)
  • Biomarkers are mainly experimental/investigational
  • Proteomic analysis of plasma or urine polypeptides (according to UpToDate, IL-2 receptor-alpha, tumor necrosis factor receptor-1, interleukin-8, and hepatocyte growth factor)
  • Flow cytometry looking for activated (CD30 positive) CD8+ T cells

Management of GVHD

Again from Yeh & Deeg (2021), this is the briefest possible summary of options which CICM trainees should be aware of, but not necessarily confident with: 


  • Methotrexate
  • Calcineurin inhibitors, eg. tacrolimus or (ergh) cyclosporine
  • Mycophenolate
  • Sirolimus
  • Cyclophosphamide
  • T-cell depletion with antithymocyte globulin (ATG)
  • Alemtuzumab (monoclonal antibody against CD52,  kills T cells)
  • Depleting T cells from the peripheral blood stem cell graft (pre-grafting)

Management of acute GVHD:

  • First line: corticosteroids (2 mg/kg/day of methylprednisolone)
  • Second line (in case of steroid treatment failure): anything goes, it seems?
    • Antithymocyte globulin (ATG)
    • Etanercept (soluble TNF receptor blocker)
    • Ruxolitinib (a Janus kinase inhibitor)
    • Pentostatin (chemo drug, adenosine deaminase inhibitor)
    • Alpha-1-antitrypsin
    • Extracorporeal photopheresis

Management of chronic GVHD

  • Topical steroids for skin involvement
  • Systemic steroids
  • Mycophenolate, cyclosporin, azathioprine, thalidomide or hydroxychloroquine as steroid-sparing agents
  • Second line agents include tacrolimis, rituximab, methotrexate or any of the second-line options used for acute GVHD.


Choi, Sung W., John E. Levine, and James LM Ferrara. "Pathogenesis and management of graft-versus-host disease." Immunology and Allergy Clinics 30.1 (2010): 75-101.

Dwyre, D. M., and P. V. Holland. "Transfusion‐associated graft‐versus‐host disease." Vox sanguinis 95.2 (2008): 85-93.

Smith, Douglas M., et al. "Liver transplant-associated graft-versus-host disease." Transplantation 75.1 (2003): 118-126.

Sanber, Khaled, Bipin Savani, and Tania Jain. "Graft‐versus‐host disease risk after chimeric antigen receptor T‐cell therapy: the diametric opposition of T cells." British journal of haematology 195.5 (2021): 660-668.

Reddy, Pavan, and James LM Ferrara. "Immunobiology of acute graft-versus-host disease." Blood reviews 17.4 (2003): 187-194.

Billingham, Rupert E. "The biology of graft-versus-host reaction." Harvey Lect. 62 (1966): 21-78.

Przepiorka, Donna, et al. "Risk factors for acute graft-versus-host disease after allogeneic blood stem cell transplantation." Blood, The Journal of the American Society of Hematology 94.4 (1999): 1465-1470.

Toubai, Tomomi, Yaping Sun, and Pavan Reddy. "GVHD pathophysiology: is acute different from chronic?." Best Practice & Research Clinical Haematology 21.2 (2008): 101-117.

Yeh, Albert C., and H. Joachim Deeg. "Graft Versus Host Disease: Management of Acute and Chronic Disease." Contemporary Bone Marrow Transplantation (2021): 365-393.

Filipovich, Alexandra H., et al. "National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: I. Diagnosis and staging working group report." Biology of blood and marrow transplantation 11.12 (2005): 945-956.

Jagasia, Madan H., et al. "National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: I. The 2014 Diagnosis and Staging Working Group report." Biology of Blood and Marrow Transplantation 21.3 (2015): 389-401.