Chapter 28. Graft-versus-Host Disease

Approximately 50,000 hematopoietic stem cell transplantation (HCT) procedures are performed worldwide each year for an expanding array of hematologic malignancies and marrow failure syndromes, metabolic disorders, and immunodeficiencies. HCT may utilize autologous, syngeneic, or allogeneic donor hematopoietic stem cell (HC). During autologous transplantation the patient's own HC are returned to the patient following preparative chemotherapy. Syngeneic transplantation is the transfer of HC between identical twins. Allogeneic HCT (allo-HCT) is the transfer of HC from a related (nonidentical) or unrelated donor to a recipient. Graft-versus-host disease (GVHD) is the primary cause of nonrelapse-related morbidity and mortality in allo-HCT and also rarely occurs following transplantation of solid organs or transfusion of blood products.

Transplantation regimens have advanced rapidly since the first successful allo-HCT was performed in 1968.1 Peripheral blood, rather than bone marrow, is now the primary source of donor HC at many transplant centers, and reduced intensity (nonmyeloablative) conditioning has permitted older patients and others who would not tolerate myeloablative chemotherapy a chance for cure with HCT. More recently, umbilical cord blood has gained prominence as a stem cell source in both pediatric and adult HCT. Donor leukocyte infusions (DLI), the administration of additional donor HC to the recipient weeks or even months after HCT, are also frequently utilized to augment graft-versus-malignancy effect.

These evolving trends in HCT, in conjunction with other known donor/recipient risk variables (Box 28-1), contribute to a wide range of reported GVHD incidence. Nevertheless, the degree of HLA-mismatch between donor and recipient remains the single most important predictor of GVHD.2 Acute GVHD develops in approximately 40% of fully matched sibling donor HCT, whereas 80% of mismatched unrelated HCT result in acute GVHD.3,4 Risk estimates of chronic GVHD also vary widely and confounding factors such as improving early posttransplant survival may be influencing the apparent trend in increasing chronic GVHD incidence.5 The most significant additional risk factor for chronic GVHD is a history of antecedent acute GVHD.5

Skin involvement is often the first indicator of acute GVHD (81%), followed by gastrointestinal (54%) and liver disease (50%).6 Similarly, the majority of patients who develop chronic GVHD manifest skin symptoms at some point in their disease course. The risk of chronic skin involvement is increased by the use of peripheral blood HCT (PB-HCT) compared with bone marrow HCT (BM-HCT). At one center, approximately 90% of patients who developed chronic GVHD following PB-HCT manifested skin symptoms.7 In most published reports, the incidence of sclerotic versus nonsclerotic chronic skin manifestations has not been differentiated. Although sclerotic involvement is less common than “lichenoid” GVHD and tends to occur later post-HCT, sclerotic features, particularly deep-seated fascial changes, may have an insidious onset, and “lichenoid” involvement is not a prerequisite to the development of sclerotic features. In one series of 196 patients post-HCT, only 7 (3.6%) developed sclerotic manifestations (mean 2.0 years after HCT).8 In a review of 133 patients who survived at least 4 months after allo-HCT, the 5-year cumulative incidence of sclerotic GVHD was 10.5% (15.5% among patients with chronic GVHD). In this series, only 21% manifested “lichenoid” changes prior to the onset of skin sclerosis.9 In a referral setting for patients with primarily refractory disease at the NIH, 81/110 (74%) consecutive patients had skin involvement, 58/110 (53%) of whom had sclerotic manifestations.10

In 1966, Billingham proposed three basic requirements for GVHD: (1) immunocompetent transplanted cells, (2) host antigens recognizable by the transplanted cells and lacking in the donor, and (3) a host incapable of mounting an immune response to the transplanted cells.11 The immunocompetent cells are now known to be T-cells, which target human leukocyte antigens (HLAs) expressed on host tissues. GVHD still develops in 40% of recipients of HLA-identical grafts, however. In this setting, GVHD is due to mismatch of key minor histocompatibility antigens (e.g., HY, HA-3).12 Tissue damage from the recipient's underlying disease, infection, and pretransplant conditioning also plays a key role in induction of the inflammatory response through pro-inflammatory cytokine production and antigen-presenting cell (APC) activation.13,14

Following activation of host APCs, T-cell activation and differentiation drives the response in acute GVHD. This appears to be primarily a Th1-driven process with massive release of interferon-γ, interleukin-2 (IL2), and TNF-α.14 Genetic polymorphisms in tumor necrosis factor (TNF)-α, interleukin-10, interferon-γ, and transforming growth factor (TGF)-β have been linked to increased risk and severity of GVHD.15–17 Although many therapies for acute GVHD target IL-2 or its receptor (CD25), these approaches (calcineurin inhibitors, daclizumab) may have the unintended consequence of adversely impacting the CD4+CD25+ regulatory T-cell population.14,18 Decreased T-regulatory cells are associated with severity of acute GVHD and poor response to GVHD treatment.19 The final effector phase of acute GVHD is characterized by cell damage via cytotoxic T-cells, natural killer cells, and soluble inflammatory mediators, including TNF-α, interferon γ, and interleukin-1.14

In comparison to acute GHVD, the pathophysiology of chronic GVHD is less well understood. Features of alloimmunity and autoimmunity and the broad spectrum of disease manifestations implicate multiple immunological pathways beyond T-cell alloreactivity. In fact, in contrast to acute GVHD, T-cell depletion of the graft does not necessarily reduce the incidence of chronic disease.20 Murine models of GVHD have demonstrated both Th1 and Th2 responses, depending on the setting; however, these models typically demonstrate specific aspects of GVHD, but do not recapitulate the full breadth of immunological and clinical abnormalities seen in human disease.21

The role of B-cell function in chronic GVHD has garnered renewed interest following the success of the anti-CD20 antibody rituximab in chronic GVHD.22 Autoantibody formation (e.g., antinuclear antibody, anti-ds DNA antibody) is a frequent finding in chronic GVHD, although the antibodies lack the specificity of typical autoimmune disease. B-cells may play several key roles in facilitating the T-cell response in chronic GVHD. Acting as APCs, B-cells prime T-cells to respond to minor histocompatibility antigens (mHA), and high-titers of antibodies directed against mHA are associated with cGVHD.23,24 Similarly, soluble levels of B-cell activating factor of the TNF family (BAFF), a cytokine which inhibits apoptosis of B-cells and promotes differentiation into plasma cells, correlate with cGVHD activity. 25,26

The mechanisms responsible for chronic GVHD-induced fibrosis in the skin and elsewhere (e.g., bronchiolitis obliterans) remains uncertain. A two-phase model has been proposed in which the innate pathway is activated through toll-like receptors, leading to an alloreactive T-cell response. This is followed by a fibrotic phase driven by platelet-derived growth factor (PDGF) and PDGF receptor (PDGFR), which in turn activates TGF-β.21 TGF-β is a potent profibrotic cytokine, capable of stimulating collagen production, abrogating metalloproteinase activity, and sensitizing fibroblasts to a constitutive-activated state via autocrine signaling.27 Furthermore, stimulatory antibodies directed against PDGFR have been identified by one group in patients with chronic GVHD as well as patients with systemic sclerosis.28,29 This has led to significant interest in imatinib mesylate, a multikinase inhibitor with potent activity against PDGFR signaling (and other receptors), for the treatment of GVHD-related fibrosis.30,31 However, to date, detection of PDGFR antibodies in sclerotic skin disease has not been replicated by other groups,32 and administration of imatinib prior to the onset of GVHD does not appear to eliminate the risk of developing skin sclerosis.33 The mechanism of action of imatinib therefore, remains unclear, and other mechanisms, including T-cell inhibition34 and inhibition of fibrosis via “nonclassic” pathways downstream of TGF-β, such as cellular Abelson (c-Abl) may be relevant.27

History

Accurate diagnosis of acute GVHD requires clinicopathologic correlation. Because the skin eruption (and histology) may be nonspecific at the time of first presentation, a careful history is invaluable. Key donor/recipient characteristics include degree of HLA-match, use of related versus unrelated donor, and T-cell depletion of the graft. Reduced-intensity conditioning may delay the onset of acute GVHD symptoms beyond the 100-day period.35 The timing of neutrophil engraftment, new medication exposures, and evidence of other organ involvement (e.g., elevated total bilirubin, diarrhea) provide additional data for clinicopathologic correlation.

Features of acute GVHD following recent blood transfusion should raise concern for transfusion-associated GVHD (TA-GVHD). TA-GVHD is an often fatal sequelae of administration of cellular blood products to immunocompromised HCT recipients, and therefore all blood products in these patients are now irradiated. TA-GVHD may also occur following transfusion of unirradiated blood products to children with congenital immunodeficiency, including Wiskott—Aldrich and ataxia-telangiectasia, as well as in the immunocompetent setting. In the latter scenario, the diagnosis may be easily missed. TA-GVHD in the immunocompetent setting follows transfusion of an unirradiated blood product that contains donor lymphocytes that are homozygous for the HLA haplotype of the recipient. A history of blood product transfusion from a relative or genetically similar population is an important feature. For example, in Japan, the estimated risk of randomly receiving blood from a homozygous donor is 1 in 874.36 In this form of TA-GVHD, the donor lymphocytes in the blood product are not recognized as foreign, leading to a GVHD reaction similar to classic acute GVHD. Beginning 10 days after transfusion, fever and skin rash (histologically consistent with GVHD) develops, followed by liver dysfunction and diarrhea. Death from pancytopenia usually occurs within several weeks.37

As with acute disease, a new diagnosis of chronic GVHD is best made based on history, cutaneous examination, and histology. A previous history of acute GVHD is the single greatest risk factor for chronic disease. Because acute symptoms may develop after 100 days posttransplant and chronic symptoms may develop before then, the revised classification of acute and chronic GVHD symptoms includes additional subtypes of GVHD with overlapping features or timing of acute and chronic symptoms (Fig. 28-1).38 Recent tapering of immunosuppressant medication or DLI given to augment the graft-versus-malignancy response are two common triggers of skin activity. DLI, in particular may present with an acute GVHD skin eruption consistent with acute GVHD rather than the papulosquamous eruption of chronic disease (Fig. 28-2D). Cutaneous or systemic infection may also induce a flare of skin GVHD, as will drug exanthems, which can result in a diagnostic challenge given the clinical and histologic similarities between viral exanthem, drug eruption, and GVHD.39 Important clues to sclerotic and fascial disease includes a history of edema of an extremity, muscle cramping, decreased flexibility, and complaints of skin tightness, particularly at the waistband and brassiere-line.10 Although GVHD in other organ systems may not necessarily flare in synchrony with skin involvement, the presence of other organ system involvement is helpful when the cutaneous features are nondiagnostic. Common GVHD symptoms include oral and ocular sicca and oral pain, particularly with spicy foods. Also common, but less specific, are symptoms of fatigue, poor appetite, and weakness. Dysphagia may indicate the presence of esophageal strictures or webbing. Bronchiolitis obliterans manifests as dry cough, wheezing, and dyspnea, but requires pulmonary function tests and computerized tomography (CT) scans to rule out infection and other etiologies. Finally, it is important to remember that despite the phenotypic variability in chronic GVHD of the skin, not every skin manifestation in a patient after HCT is due to GVHD, so a careful dermatologic history to detect other possible diagnoses is prudent.

Cutaneous Lesions

Acute GVHD initially presents with erythematous-dusky macules and papules of the volar and plantar surfaces and ears that may rapidly become a diffuse morbilliform exanthema (Fig. 28-2A and 28-2B; Box 28-2). Very early involvement may manifest as erythema limited to hair follicles (Fig. 28-2C). Pruritus is variable and is not useful to distinguish acute GVHD from other causes. Erythroderma may develop, and, in severe cases, spontaneous bullae with skin sloughing resembling toxic epidermal necrolysis. Widespread erythrodermic involvement, particularly the presence of skin sloughing portends a very poor prognosis. In contrast to chronic disease, postinflammatory pigmentary changes following acute GVHD are uncommon.

In appreciation of the tremendous variability in clinical presentation of chronic skin GVHD, it is no longer useful to dichotomize chronic GVHD of the skin into either “lichenoid” or “sclerodermoid” categories. In the transplant community, the term “lichenoid” has been utilized to denote any involvement of the skin in which erythema or scaling is present; however, “lichenoid” is a histologic pattern, not a clinical one, and, therefore, usage of it is best reserved to pathologic description. Futhermore, although chronic GVHD may resemble lichen planus (Fig. 28-3), other patterns are frequently observed, such as poikiloderma (Fig. 28-4) and skin lesions resembling lupus erythematosus, keratosis pilaris, or psoriasis.40 Postinflammatory hyperpigmentation is common following the resolution of epidermal involvement, particularly in darkly pigmented individuals, and may persist for many months after the skin disease becomes quiescent.

The fibrotic changes of chronic GVHD are also remarkably variable, and the term “sclerodermoid” is an inadequate descriptor of the varied sclerotic tissue abnormalities in the dermis, subcutaneous tissue, and fascia (Fig. 28-5). As in systemic sclerosis, an edematous phase may herald the onset of skin fibrosis, but fingers and toes are usually spared and the typical acral to proximal progression characteristic of systemic sclerosis is not seen in chronic GVHD. In constrast to systemic sclerosis, facial involvement is rarely involved in sclerotic-type GVHD. As mentioned above, fibrosis may occur primarily in the upper dermis, through the full-thickness of dermis, or in the subcutaneous fat and fascia. Early superficial fibrotic involvement resembles lichen sclerosus, often manifesting as porcelain-white atrophic plaques on the upper back (Fig. 28-5A). A common pattern of GVHD-associated fibrosis involves patchy sclerotic plaques with hypo- and hyperpigmentation mimicking morphea. Sclerosis of this type may exhibit an isomorphic response, localizing to the sites of minor skin trauma, particularly the waistband area, or may develop at sites of previous scar formation (Fig. 28-5B).41 Diffuse dermal involvement may result in a “pipe-stem” appearance of the lower extremities with marked reduction in limb volume and overlying shiny hidebound skin with loss of hair resembling scleroderma (Fig. 28-5C). Deeper involvement of the subcutaneous fat results in irregular hyperpigmented sclerotic plaques with intervening areas of edematous skin closely resembling deep morphea/morphea profunda.42 Bullae may develop at sites of fibrosis, particularly on the lower legs, as a result of dermal edema, as has been described in bullous morphea profunda.43 Patchy hyperpigmentation (“leopard spots”) may be visible prior to the diagnosis of dermal sclerotic involvement.44

Primary involvement of the subcutaneous fat and fascia results in a diffuse firm, rippled pattern to the skin resembling eosinophilic fasciitis (Fig. 28-5D).45 Features of overlying epidermal GVHD involvement and pigmentary changes may be absent. Fascial involvement is often most visible on the medial arms and thighs and be accentuated by abduction and supination of the arm. Prominent “grooving” demarcating fascial bundles and along the path of superficial vessels may be observed. Careful palpation of the skin is helpful in detecting deep-seated irregularities in skin texture and differentiation from cellulite. Dermal fibriosis or fascial involvement without overlying dermal thickening may lead to progressive loss of joint range of motion and contracture formation.

Nail involvement in chronic GVHD typically results in longitudinal ridging and thin, easily broken nails. Partial or complete anonychia and dorsal pterygium formation may occur. Other unusual skin sequelae of chronic GVHD include milia formation, porokeratosis, often on the buttock area,46 angioma formation at sites of skin sclerosis,47 nipple hyperkeratosis,48 vitiligo,49 and alopecia, either diffuse or focal areas of alopecia areata.50

Different manifestations of sclerotic and and nonsclerotic skin disease may be present in the same individual, making accurate quantification of disease activity challenging. The chronic GVHD NIH Consensus Development Project provided more precise terminology for organ system involvement and defined features specific for diagnosis of chronic GVHD in the setting of HCT (Box 28-3). Diagnostic cutaneous features of GVHD include poikiloderma, lichen-planus-like lesions, and sclerotic skin changes.38

Related Physical Findings

Acute GVHD is primarily a disorder of the skin, GI tract, and liver (Box 28-2), typically presenting with skin rash, new onset elevation of total bilirubin, and/or voluminous diarrhea. By contrast, chronic GVHD is remarkably diverse in its breadth of organ system manifestations (Box 28-3). The most frequently affected sites are skin and nails, oral mucosa, eyes, liver, lungs, and marrow (usually thrombocytopenia).5 Esophageal webs/strictures, vagino-vulvar disease, myositis, nephrotic syndrome, and pericarditis are less frequent sequelae of chronic disease.

Mucosal disease is second only to skin involvement in frequency in chronic GVHD. Mucoceles are common, as are erosions, lichen-planus-like changes with Wickham's striae, and sicca symptoms. Dryness and violaceous erythema of the lips are common. Genital involvement significantly impairs sexual function and quality of life and may be overlooked if a specific examination and directed questions regarding genital symptoms are not undertaken. Involvement of the penis may induce phimosis. Vulvo-vaginal involvement presents as erythema, erosions/fissures, vestibulitis, vaginal stenosis, labial resorption, or complete agglutination of the introitus leading to hematocolpos (Fig. 28-6).51

Histopathology

The histological grading scale for acute GVHD is shown in Table 28-1. The hallmark feature of acute GVHD is the presence of necrotic keratinocytes accompanied by a dermal lymphocytic infiltrate (usually sparse) and basal vacuolar alteration (Fig. 28-7). Early GVHD involvement with follicular erythema correlates with involvement limited to the hair follicle. Subepidermal cleft formation (Grade III) is indicative of more severe involvement, whereas complete separation of epidermis from dermis (Grade IV) correlates with clinical findings resembling toxic epidermal necrolysis. Grade IV involvement may be impossible to differentiate histologically from drug-induced toxic epidermal necrolysis and requires careful clinical correlation. The presence of eosinophils has been used in the past to argue against a diagnosis of GVHD; however, the presence of scattered eosinophils may lead to a false diagnosis of drug eruption52 and, unless very large numbers of eosinophils are present, this feature cannot be used as a reliable indicator of a drug hypersensitivity reaction.53 Engraftment syndrome is a poorly understood phenomenon at the time of neutrophil engraftment following autologous-HCT or allo-HCT characterized by a nonspecific erythematous skin eruption, fever, and pulmonary edema.54 Histologically, it may not be possible to distinguish engraftment rash from early (Grade I) acute GVHD.

Epidermal changes in chronic GVHD may be indistinguishable from those of acute disease (Fig. 28-8A). Acanthosis and wedge-shaped hypergranulosis may be seen. Sclerotic involvement of the upper dermis may resemble lichen sclerosus, with atrophy, hyperkeratosis, follicular plugging, and pale, homogenized appearance of the upper dermis collagen (Fig. 28-8B).45 If epidermal changes of GVHD are not present, dermal fibrosis with thickened collagen bundles and loss of periadnexal fat involvement may be indistinguishable from morphea/scleroderma. Subcutaneous and fascial involvement accordingly demonstrates changes in the fat septae and fascia, including thickening, edema, and fibrosis. Variable lymphocytes, histiocytes, and eosinophils may be seen.45

Histology of involvement of the oral mucosa reflects similar interface changes as those seen in epidermal GVHD, but without associated acanthosis.55 Lymphocytic infiltration of the salivary glands resembles changes seen in Sjögren's syndrome.

Diagnosis of acute GVHD skin involvement is based on histopathologic correlation, particularly exclusion of drugs and infectious causes. The presence of a normal leukocyte count is indicative of engraftment but no specific laboratory testing is diagnostic. Liver function testing and total bilirubin levels and quantification of diarrhea volume are used in conjunction with skin disease to stage the disease (Table 28-2).

Although autoimmune markers are seen in the majority of patients after alloHCT, their presence is generally not specific for the development of chronic GVHD manifestations, with the possible exception of sclerotic disease. In one study, elevated ANA titer was detected in 70% of patients with limited chronic disease and 94% of patients with extensive chronic disease compared to 23.5% of patients who did not develop chronic GVHD.56 The presence of more than one autantibody also correlated with risk of extensive disease (p = 0.04); however, ANA titer does not correlate with disease severity. In this study, the presence of a nucleolar ANA pattern also indicated a potential association with sclerotic disease (p = 0.06).56 In another multivariate analysis limited to sclerotic-type chronic GVHD patients, the presence of autoantibodies and serum eosinophilia were both associated with increased risk of sclerotic-type chronic GVHD.9

Identifying specific biomarkers of disease activity is an area of research emphasis in acute and chronic GVHD.57 Plasma levels of elafin, a protease secreted in response to IL-1 and TNF-α, was recently identified as a candidate marker capable or differentiating acute GVHD skin from rashes of other etiologies. In this study, immunohistochemical staining of skin biopsies for elafin also discriminated acute GVHD from drug exanthem, suggesting a potential diagnostic application of this biomarker.58

Special Tests (Including Imaging Studies)

Suspicion of subcutaneous sclerotic and fascial disease and myositis may be confirmed by magnetic resonance imaging, particularly in cases in which definitive sclerotic changes are not observed or when a fascial or muscle biopsy is deferred.45,59,60

See Box 28-4.

Complications

Skin erosions and ulceration due to chronic GVHD may lead to secondary infection. Sclerotic changes resulting in restriction in joint function lead to functional disability and joint contractures. Restrictive lung disease may result from sclerotic involvement of the torso. HCT survivors are at increased risk for melanoma61 and nonmelanoma62 skin cancer due to previous exposure to ionizing radiation, GVHD-associated immunodysregulation, and immunosuppressive treatment for GVHD. The risk of cutanous squamous cell carcinoma (SCC) may also be increased by long-term treatment with voriconazole, a potent photosensitizer, which may be employed for antifungal treatment or prophylaxis.63 Multiple SCC have also been reported after PUVA for GVHD.64

Although the presence of GVHD is associated with decreased risk of malignancy relapse, GVHD is also a cause of significant morbidity and mortality, particularly in patients who develop refractory disease. A number of systemic risk factors portend a poor prognosis, including a history of progressive involvement from acute to chronic GVHD,65 thrombocytopenia (fewer than 100,000 cells/mL),66 elevated bilirubin,67 older age, gastrointestinal symptoms, and lack of response to therapy at 6 months.68 The two primary dermatologic features associated with poor prognosis are extensive (>50%) skin involvement69 and “lichenoid” skin histology.65

Management of Acute GVHD

Treatment of acute GVHD is usually undertaken in the hospital, given the proximity to the date of HCT and the need for close observation. Patients with mild (Grade I) skin involvement without hepatic or gastrointestinal symptoms may respond to high-potency topical steroids. However, more severe skin involvement or the presence of internal organ involvement necessitates treatment with systemic corticosteroids (methylprednisone 2 mg/kDa/day). Patients with skin sloughing require meticulous skin care, infection surveillance, and fluid management similar to toxic epidermal necrolysis. Approximately 50% of patients respond to systemic corticosteroids—however, those who require salvage therapy typically receive one or more immunosuppressive agents, including calcineurin inhibitors (tacrolimus, cyclosporine), mycophenolate mofetil, and sirolimus, which are of variable success (Box 28-5).70 Phototherapy (PUVA,71 NB-UVB,72 UVA173) has also been used in small series for acute GVHD, but is logistically challenging in the inpatient setting and should be administered cautiously to avoid inducing erythema.

Extracorporeal photopheresis (ECP), anti-TNF–α therapy, and multipotent mesenchymal stromal cells (MSC) are additional strategies that have shown recent success for the treatment of acute skin GVHD. In a review of salvage therapies for acute GVHD, 60%–76% of patients with skin involvement responded to ECP; however, responses decreased with increasing skin severity.70 Levine et al74 demonstrated complete remission (CR) of skin symptoms in 81% of patients treated with steroids and etanercept compared to steroids alone (CR = 47%). Similarly, infliximab has also shown variable success in acute treatment-refractory skin GVHD (33%–60%).70 Finally, preliminary reports of the success of MSC, bone marrow-fibroblast derived cells capable of differentiation into adipocytes, chondrocytes, and osteoblasts, in patients with refractory acute GVHD has generated significant interest in this novel therapy.75–77 In a 2008 study, 39/55 (71%) of participants with steroid-resistant acute GVHD sustained a complete or partial response to MSC infusion.77 Responses were seen regardless of MSC source (HLA-matched, haploidentical, or third party unmatched donors), and immunogenicity was not observed. The immunomodulatory mechanism of MSC is unclear, but may be through induction of regulatory T-cells.78,79 Several MSC studies are underway for the treatment or prophylaxis of acute and chronic GVHD as well as for other chronic conditions, including Crohn's disease, multiple sclerosis, systemic sclerosis, and lupus erythematosus.

Management of Chronic GVHD

Among the myriad topical, phototherapy-based, and systemic treatments that have been used in patients with chronic GVHD who cannot be tapered from systemic corticosteroids or who are steroid-refractory, no single treatment has demonstrated proven superiority (Table 28-3). Determination of a preferred second-line agent has been complicated by poor understanding of the disease process and a lack of high-quality clinical trials. The need to spur clinical trial development in the field of chronic GVHD was acknowledged by the chronic GVHD NIH Consensus Project, which included a standardized system of organ system assessments and recommendations for clinical trial design80 Unfortunately, validated measures of cutaneous disease activity are still lacking, and the most common skin assessments tools, body surface estimates and use of Rodnan scoring (derived from systemic sclerosis trials) are not applicable for all manifestations of chronic GVHD skin activity.81 Ideally, dermatologic collaboration in future therapeutic trials will permit better quantification of cutaneous disease response.

The dermatologist should play a key role in the multidisciplinary approach to chronic GVHD management, beginning with careful assessment of the subtype and extent of skin involvement. Together with an understanding of other organ system activity, infection risk, relapse risk, and GVHD prognostic risk factors, a decision regarding the appropriateness of topical, physical (e.g., phototherapy), and systemic therapy can then be made. If systemic therapy is prescribed by the transplant physician, periodic dermatologic monitoring is advised to differentiate adverse drug reactions or other new skin disease from GVHD,82 to assess cutaneous disease response, and to monitor for infection and skin malignancy.

Nonsclerotic lichen-planus like and other papulosquamous chronic GVHD manifestations may respond well to topical steroid treatment and serve to reduce exposure to systemic immunosuppression.83 Topical emollients and antipruritic agents may provide relief of pruritus and skin irritation; however, oral antihistamines may worsen sicca symptoms in patients with oral and ocular dryness. Choi and Nghiem84 described a response to topical tacrolimus 0.1% ointment in 13/18 patients with chronic GVHD; however, all patients eventually required other therapy to control their skin disease. Subsequent reports have also described response to topical pimecrolimus 1% cream.85,86 Topical calcineurin inhibitors are particularly useful for treatment of areas at high risk of skin atrophy, such as the face (including the lips) and intertriginous surfaces. Topical tacrolimus may not be tolerable at sites of significant inflammation or erosions. Hydroquinone in combination with tretinoin and topical dexamethasone has anecdotally been reported to improved periocular lichenoid type chronic GVHD and hyperpigmentation.87 Topical tretinoin may also benefit milia formation following GVHD skin activity.

Phototherapy may be of benefit for both sclerotic and nonsclerotic chronic GVHD, but data are limited to anecdotal cases and a small number of noncontrolled case series. Vogelsang88 described improvement in 31/40 patients treated with PUVA.88 Three patients in this series had skin sclerosis—two demonstrated transient benefit, but developed severe phototoxicity, and the third did not respond to treatment. Smaller series with PUVA-bath (6 patients),89 narrow-band UVB (10 patients, pediatric),90 and UVB (5 patients),91 have also described chronic GHVD responses, primarily in patients with “lichenoid” disease. Over the last several years, however, there has been growing experience with UVA1 for sclerotic skin conditions, suggesting potential application in chronic GVHD. Longer wavelength UVA1 (340–400 nm) does not require psoralen ingestion/topical application and penetrates deeper into the dermis than full spectrum UVA. Several reports have described skin softening following UVA1 treatment of lichen sclerosus,92 localized morphea,93–95 and sclerotic-type GVHD.73,96–98 Wetzig et al73 used medium-dose UVA-1 phototherapy in seven patients with lichenoid GVHD and three with sclerotic GVHD. All three patients with sclerotic GVHD demonstrated partial response or improvement. Ständer et al97 described softening of skin lesions, improved joint mobility, and healing of skin erosions in five adult patients with medium-dose UVA-1 and one child treated with low-dose UVA-1. Calzavara Pinton et al98 described five patients with sclerotic involvement treated with MD UVA-1 therapy leading to complete resolution in three patients and partial response in two patients. UVA-1 may accentuate pigmentary abnormalities.99 Although UVA-1 is not yet widely available in the Unites States, it appears to be well tolerated, acceptable for pediatric use,96 and is not associated with persistent photosensitivity or potential gastrointestinal issues that may occur with oral psoralen use.

Phototherapy may be appropriate for patients with limited epidermal or sclerotic disease in whom systemic therapy is not otherwise warranted (e.g., without internal organ system involvement), or in whom systemic immunosuppressive therapy is contraindicated (e.g., active infection); however further controlled trials are needed to directly compare phototherapy modalites and to determine the optimum dose and treatment schedule. Skin cancer risk assessment and concurrent use of photosensitizing medications should also be considered. Multiple squamous cell carcinomas have been reported following PUVA treatment for chronic GVHD 64 and the risk of melanoma is elevated in patients following HSCT.61 Photosensitizing medication use is common, including voriconazole therapy, which may further increase the risk of squamous cell carcinoma formation in the setting of chronic GVHD.63

Extracorporeal photopheresis (ECP) is another option for patients with cutaneous disease, particularly patients with extensive or sclerotic involvement. During ECP, the white cell compartment of the blood is removed from the patient via pheresis, mixed with 8-methyoxypsoralen, irradiated with UVA light, and then returned to the patient. In a retrospective review of 71 chronic GVHD patients who were treated with ECP, 59% of patients with cutaneous involvement responded, including 67% of those patients categorized as sclerotic involvement.100 ECP may be particularly useful for patients with deep-seated sclerotic involvement of the subcutaneous tissue and fascia. Although GVHD-related fasciitis resembles eosinophilic fasciitis (EF), in contrast to EF, it does not respond well to steroid therapy and may result in significant long-term functional disability. Several case reports describe successful use of ECP for EF,101 and GVHD-related fasciitis.41,102,103 ECP is a time-consuming procedure and requires a dedicated pheresis center which is not available at all medical facilities. As with phototherapy, the optimal frequency and duration of ECP treatment is unclear. Typically, intensive treatment (2× weekly or every other week) is initiated, followed by an attempt to decrease frequency if a response is achieved.

Limited data are available supporting the use of systemic retinoids for chronic GVHD. Marcellus et al104 reported improvement in 20/27 evaluable patients with sclerotic disease treated with etretinate; however, six patients could not tolerate the treatment due to scaling or skin breakdown. Ghoreschi et al105 described PUVA-bath treatment in 14 patients with sclerotic-type GVHD, five of whom received concurrent treatment with isotretinoin 10–20 mg/daily. Overall improvement was reported in 7/14 patients; however, skin ulceration was a significant issue in both PUVA-bath only and combination treatment groups, and the small sample size precluded statistical comparison between groups.105 Further prospective studies are needed to determine the tolerability and efficacy of systemic retinoid therapy.

Imatinib mesylate, a multikinase inhibitor with activity against bc-abl, c-kit, PDGFR, and other kinases, has been reported to benefit patients with sclerotic GVHD in a small number of case reports.30,31,106 The drug is generally well tolerated in the setting of treatment for chronic myelogenous leukemia; however, the tolerability and efficacy of the drug in chronic GVHD is an area of ongoing clinical trial investigation. Common side effects include peripheral and periorbital edema, myalgia, and fatigue. Second generation agents with similar targeted tyrosine kinase inhibitory activity (nilotinib, dasatinib) also hold potential as therapeutic options for sclerotic disease, as does the use of MSC, based on the experience with acute GVHD. Recently, a single case of sclerotic-type GVHD was reported with a response to MSC therapy.107

Treatment of Chronic Oral and Vulvo-Vaginal Disease

Limited oral mucosal disease can be controlled with application of high-potency topical corticosteroid gel (fluocinonide gel 0.05%, clobestasol gel 0.05%). Refractory lesions may respond to intralesional triamcinolone injection (0.3–0.4 mL/L cm2).83 Topical application of tacrolimus 0.1% ointment may also be used;108 however, systemic absorption has been reported109 and, therefore, serum tacrolimus levels are reasonable following initiation of intra-oral treatment. Generalized oral disease can significantly impair oral intake and quality of life and often result in the need for systemic intervention. Corticosteroid rinses (dexamethasone 0.5 mg/mL; prednisolone 15 mg/mL) are beneficial for widespread involvement and should be swished in the mouth 4–6 minutes 4–6 times daily.83 Cyclosporine and azathioprine rinses may also be used for refractory disease, but require pharmacy compounding. As mentioned above, patients with salivary gland disease should avoid oral antihistamines as well as other xerogenic medications (SSRIs, tricyclic antidepressants). Dental hygiene is very important in patients with decreased salivary function and home fluoride treatment is frequently recommended. Salivary stimulants (e.g., sugar-free gum) and sialogogue therapy (cevimeline, pilocarpine) are recommended for patients with severe salivary gland dysfunction.83 Although sclerotic involvement of perioral skin involvement is uncommon, in this setting aggressive systemic therapy is indicated.

Genital erosions and fissures associated with chronic vulvo-vaginal disease may be treated with clobetastol proprionate ointment nightly, which should be tapered to a maintenance level of 2–3 times weekly. If estrogen is not contraindicated, hormone replacement via topical cream, vaginal ring, or oral replacement may improve genital skin integrity. Limited vaginal scarring/synechiae can be treated with dilators or manual lysing; however, thick vaginal scarring may require surgical intervention.110

GVHD prevention begins prior to transplantation with the selection of the most closely HLA-matched donor, the GVHD prophylaxis regimen and, in some cases, manipulation of the T-cell content of the graft. T-cell depletion is accomplished through ex vivo T-cell negative selection or enrichment of the CD34+ stem cell population, or through in vivo treatment with anti-T-cell therapy. The benefits of T-cell depletion, however, are offset by higher rates of graft failure, cancer relapse, and infection.14 Prophylactic immunosuppressive therapy is initiated concomitantly with the administration of the hematopoietic graft, but, as with T-cell depletion, such therapy must be balanced with potential for diminished graft-versus-leukemia/lymphoma effect and long-term infection risks. In general, available strategies for the prevention of acute GVHD are rarely effective in the prevention of chronic GVHD, emphasizing the distinct pathophysiology of these two GVHD manifestations. Ideally, personalized immunogenomics will evolve to allow careful titration of T-cell graft content and prophylactic immunosuppression to maximize graft acceptance and graft-versus-leukemia effect and at the same time minimize infection risk and other complications associated with long-term immunosuppression.

Similar to solid-organ transplantation, skin cancer screening and patient education regarding photoprotective measures is a key preventive strategy in patients with chronic GVHD.83 Patients are also at elevated risk of systemic infection, and therefore, implementation of preventive infectious disease recommendations and careful monitoring for cutaneous infection, particularly in patients with chronic skin erosions/ulcerations, is prudent.83 Finally, patient education regarding early signs of skin sclerosis and fascial involvement, including skin tightness, edema, muscle cramping, and range of motion restriction, may facilitate early diagnosis and initiation of treatment.