Predictors for pulmonary mortality was determined in a cohort of 79 patients with acute-GVHD of the skin. The acute-GVHD treatment was corticosteroids and photochemotherapy (Photosensitization with oral 8-methoxysalen and Ultraviolet light type A) with or without concomitant methotrexate.

The study encompassed 79 patients who were retrospectively identified. Eligible patients were those who had developed acute-GVHD of the skin and had been treated by photochemotherapy at the dermatology department at Huddinge University Hospital before the end of 2005. The follow-up of survival and relapse was a minimum ten year follow up. Patients with elevated bilirubin or excessive diarrhoea fulfilling the criteria of acute-GVHD of liver or gastrointestinal acute-GVHD were excluded, this to prevent confounding of the primary outcome measure by secondary ARDS elicited from the viscera and to limit the confounding of generalized toxicity or infections. The patients were diagnosed in accordance with the Glucksberg criteria, i.e. the extent of skin rash was stratified into skin disease stage 1 for an erytomatoeus rash covering <25% of the TBSA, skin disease stage 2 for a rash affecting 25 - 50% of the BSA and skin disease stage 3 for a rash affecting more than 50% of the TBSA.(Glucksberg H., 1974, Ringden O., 1996). The acute-GVHD diagnosis was supported with biopsy- and post-mortem histopathology. The patients who received methotrexate i.v. as an immunosuppressive treatment combined with photochemotherapy were compared with the patients who only received photochemotherapy. Photochemotherapy was administered at the department of dermatology where treatment data, including number of treatments and dose, treatment effect and adverse effects of photochemotherapy were recorded. Methotrexate was administered at the transplant unit. Non-negotiable variables and outcomes where primarily chosen to limit the bias. Toxicity was estimated by photo toxicity, renal impairment, liver damage and myelosuppression. The effect on acute-GVHD, creatinine, ALAT, leukocyte counts was determined from the prospective data records at the transplantation unit. All data including cause of death was cross checked with the centre for allogeneic stem-cell transplantation, (CAST) quality register and the records from CAST, the intensive care unit and the department of haematology including the death certificate. The study was undertaken in accordance with the Helsinki declaration and approved by the regional ethics committee, number 2012/969-31/3 with addendum 2014/1569-32 and number 425/97.

Treatment of acute-GVHD The acute-GVHD was treated with Corticosteroids in a dose of 2mg/kg prednisolone i.v. with additional bolus doses of methylprednisolone at the hands of the attending doctor. The variable corticosteroid treatment at the start of photochemotherapy was divided into: no corticosteroids, corticosteroid treatment but not corticosteroid resistant acute-GVHD, and finally corticosteroid resistant acute-GVHD (Remberger M., 2001). Oral 8-methoxypsoralene (8-MOP), (0.4-0.8 mg/kg), was ingested 1.5-2 h before the BSA was radiated by Long-wave UVA (320-400 nm) from a Waldmann UV1000 supine unit (Waldmann, Villingen-Schwenningen, Germany) with 26 Waldmann F85 100-W fluorescent photochemotherapy lamps or a Waldmann UV3003K half-body unit with 15 Waldmann F85 100-W photochemotherapy lamps (Parrish J.A., 1974, Henseler T., 1981). During UVA the genital area of male patients was protected. Eyes were shielded for 24 h thereafter during therapy. The dose of photochemotherapy was divided into the binary variable; low dose i.e. (0 - 9) treatments versus 10 treatments or more. Methotrexate was administered i.v. in 7,5 mg/m2 body surface area (1-3) times not more often than three times a week (Nassar A 2014). Concomitant injection of methotrexate during the period of photochemotherapy was registered as the binary variable; present or not present.

Outcome The Primary outcome; Crude pulmonary mortality was defined as lethal outcome of pulmonary disease and comprised IPS including interstitial pneumonitis with or without pulmonary infection, but also pneumonia and undefined respiratory insufficiency or interstitial fibrosis. As secondary outcome pulmonary mortality was divided into a binary variable; where those causes primarily associated with severe immunosuppression e.g. pulmonary mortality secondary to opportunistic infections; i.e. CMV-pneumonitis, fungal pneumonia or with a diagnosis of CMV-infection or fungal infections at the time of pulmonary mortality, was separated from the group of patients who died from pulmonary mortality without concomitant opportunistic disease (Yanik G., 2005, Watkins T.R., 2005, Forslow U., 2006, Bjorklund A., 2007) Chronic graft-versus-host disease was both included as a secondary outcome of acute-GVHD treatment and included as a predictor in the multivariate analysis for non-opportunistic pulmonary mortality.

The study size

The Study Size was all the patients in the closed photochemotherapy cohort who had cutaneous acute-GVHD without concomitant visceral disease at the start of photochemotherapy.

Statistical methods Shapiro-wilk was used to define if the variables were parametric or non-parametric. Parametric data was described with mean and +- SD, while non-parametric data was described with median and max-min. Kaplan and Meier curves was used to depict cumulative incidence of survival and Cox proportional hazards ratio was used to evaluate the risk for death in respiratory disease not explained by infectious agents. Log-rank test was used to variables that did not fit into the cox-model.

Inclusion Criteria:

- Treatment with photochemotherapy for aGVHD of the skin at the Dermatology Department at Huddinge Hospital before the end of 2005.

Exclusion Criteria:

- Retransplantation or DLI before photochemotherapy

Transplantation at Huddinge University Hospital (Karolinska UH) and treated with photochemotherapy at the Dermatology department before the end of 2005.