Oral mucositis after tacrolimus/sirolimus or cyclosporine/methotrexate as graft‐versus‐host disease prophylaxis

Abstract Objectives To determine whether treatment with tacrolimus plus sirolimus (Tac/Sir) as a prophylaxis for graft‐versus‐host disease worsens severe oral mucositis and delays healing compared to cyclosporine plus methotrexate (CsA/Mtx) following haematopoietic stem cell transplantation. Subjects and Methods The study comprised 141 patients: 73 randomized to receive Tac/Sir and 68 to receive CsA/Mtx. The oral mucositis assessment scale and toxicity grading according to WHO were used to assess the severity, peak and duration of oral mucositis from the day −3 to day 24 post‐transplant. Results Eighty‐seven patients developed oral mucositis in the first 24 days post‐transplant. No significant difference in oral mucositis severity between the Tac/Sir and CsA/Mtx groups was observed. The peak oral mucositis score occurred on day 10 in both groups. Although oral mucositis scores had returned to baseline in the CsA/Mtx group on day 24 post‐transplant, no significant difference compared with the Tac/Sir group was found. Conclusions The introduction of tacrolimus/sirolimus as a graft‐versus‐host disease prophylaxis in haematopoietic stem cell transplantation increased neither the incidence nor severity of oral mucositis compared with cyclosporine/methotrexate. Furthermore, oral mucositis healing was not prolonged and followed the same time pattern as cyclosporine/methotrexate.

Continual advancement in transplant procedures has steadily improved patient outcome over time, but graft-versus-host disease (GVHD) remains a serious complication of HSCT. Both acute and chronic GVHD contribute significantly to morbidity and mortality after treatment (Deeg, 2007;Gooley et al., 2010;Remberger et al., 2011;Ziakas et al., 2014).
Oral mucositis (OM) is a frequent and sometimes severe side effect of post-transplant conditioning therapy that refers to painful mucosal ulcerations of the mouth and gastro-intestinal tract (Garming Legert et al., 2014). Between 75% and 100% of HSCT recipients develop OM (Elad et al., 2015). Direct cell injury mediated by chemotherapy and radiation characterize OM. Cell injury is a consequence of a complex cascade beginning with cell death and the release of reactive oxygen species, progression through a series of steps in which biological pathways are activated and amplified, and culminating in mucosal ulcerations .
Since the 1980s, a combination of cyclosporine and methotrexate (CsA/Mtx) has reduced severe acute GVHD (aGVHD) and improved treatment-related mortality after HSCT (Reshef, 2012;Ringdén et al., 1993;Storb et al., 1989). In the last decade, a regimen of tacrolimus and sirolimus (Tac/Sir) has shown promising immunosuppressive capacity and other desirable properties in solid organ transplantation (Cutler et al., 2004(Cutler et al., , 2007. Sirolimus binds to the mammalian target of rapamycin (mTOR) in T cells and suppresses T-cell proliferation by inhibiting progression from the G1 to the S phase of the cell cycle (Li et al., 2014). Sirolimus has anti-proliferative effects on fibroblasts, endothelial cells and smooth muscle cells (Akselband et al., 1991), suggesting impaired mucosal healing (Macdonald, 2007). Recently, studies on sirolimus in HSCT have found a decreased incidence of aGVHD and treatment-related toxicity at the expense of higher rates of transplant-associated thrombotic microangiopathy and related endothelial injury syndromes (Cutler et al., 2014;Shayani et al., 2013). Aside from HSCT, a high incidence of painful oral ulcers has been reported in the kidney transplant setting (van Gelder et al., 2003).
These efforts to address GVHD, including new combinations of immunosuppressive drugs and new immunosuppressive strategies, confer a risk of new side effects. One risk is the development of oral lesions in patients subjected to HSCT; another is that healing of OM lesions may be prolonged. In severely neutropenic patients, higher risks of generalized infections and prolonged hospital stays are possibilities (Garming Legert et al., 2014;Sonis et al., 2001).
Our group has previously published the results of a prospective randomized trial comparing the immunosuppressive regimens of Tac/Sir with CsA/Mtx after HSCT (Törlén et al., 2016). In that trial, no significant differences could be demonstrated between the two prophylactic regimens when comparing the cumulative incidence of aGVHD with transplant-related mortality.
The aim of the present study, a substudy of the previous clinical trial, was to determine whether treatment with Tac/Sir as prophylaxis for GVHD worsens severe OM and delays healing compared to the standard regimen of CsA/Mtx.

| Patients
In the main clinical trial, patients were randomized in a prospective, open-label, phase III, multicentre trial comparing Tac The study followed the guidelines of the Declaration of Helsinki. All authors attest the accuracy of the reported study data and adherence to the study protocol.
The primary endpoint of the clinical trial was NIH grades II-IV aGVHD in the two treatment groups within 200 days post-HSCT.
The secondary endpoint was incidence and severity of OM in enrolled patients. All eligible participants, or their parents or guardians if the participant was under 18 years of age (see Törlén et al., 2016 for a detailed description of the inclusion and exclusion criteria), signed informed-consent forms.
The present OM substudy comprised only the Stockholm cohort (141 patients, Table 1): 73 patients were randomized to receive Tac/ Sir, and 68 were randomized to receive CsA/Mtx as prophylaxis for GVHD. doses were previously described (Törlén et al., 2016). Stem cell source was peripheral blood progenitor cells or bone marrow.

| Haematopoietic stem cell transplantation treatment protocols
Supportive care followed previously published institutional standards (Forslöw et al., 2006). For all diagnoses, patients in the CsA/Mtx group started CsA on day −1 and Mtx 15 mg/m 2 on day + 1, with consecutive doses of 10 mg/ m 2 on days + 3, +6 and +11. Following HSCT, both regimens were discontinued after tapering 3 to 6 months, depending on grafts and in the absence of GVHD.

| Assessment of acute GVHD
The attending physicians diagnosed aGVHD clinically and assigned a grade from 0 to IV using previously published criteria (Przepiorka et al., 1995). Biopsies from the skin, gut and liver were assessed according to the routines at the centre.

| Oral examinations
Baseline examination of the oral cavity included a radiographic and a clinical examination approximately 14 days before HSCT, before the start of conditioning treatment. If needed, infectious foci in the oral cavity were treated conservatively. The nurses were trained and calibrated to diagnose OM according to the

International Classification of Diseases from the World Health
Organization (World Health Organization, 1979). Post-treatment, the nurses conducted oral examinations of the patients daily, from 3 days before the transplantation, until day 24 after HSCT or discharge from the hospital (n = 141). The patients self-reported pain in the oral cavity on a visual analogue scale. A dentist visited the patients 3 days a week to examine the oral mucosa and record the clinical features of OM using the oral mucositis assessment scale (OMAS; Sonis et al., 1999) and the World Health Organization Oral Toxicity Scale Grading of Oral Mucositis (Sonis et al., 2001). There was a good correlation between the WHO OM grading scale and the OMAS (r = 0.74, p < .001).

| Standard oral hygiene protocol
The standard oral hygiene protocol included careful tooth brushing twice a day with interdental cleaning. All patients were instructed to suck on ice chips throughout conditioning, up to the day of transplantation, and especially during chemotherapy. Patients with symptoms of OM were recommended to continue sucking ice chips as much as possible. They were also instructed to rinse their mouth with saline solution once every waking hour-from transplantation until the neutrophil blood cell count exceeded 0.5 x 10 9 /litre. Patients with symptoms of dry mouth received a saliva substitution or lubricants.

| D ISCUSS I ON
The present study found that, as a GVHD prophylaxis of OM, Tac/ Sir following HSCT increased neither the incidence nor severity of OM compared with CsA/Mtx. Furthermore, OM healing was not prolonged and followed the same time pattern as CsA/Mtx. In all, 62% of the patients who underwent HSCT were diagnosed with OM, and MAC was an important risk factor. In a study on a similar HSCT population, Shouval et al. (2019) found that the incidence of moderate-to-severe OM was as high as 83% and depended primarily on conditioning intensity and GVHD prophylaxis. The literature describes several other risk factors in the HSCT setting, often with contradictory clinical results (Barasch & Peterson, 2003).
OM (WHO grades 2-4) was diagnosed in 58% of patients receiving Tac/Sir compared to 64% receiving CsA/Mtx, a non-significant difference.
The present study found that the peak incidence of OM occurred on day 10 after HSCT and that the WHO OM score had returned to baseline on day 24. This corresponds to previous studies reporting a peak after 10-11 days and complete healing after 21-23 days (Cutler et al. 2005;Garming Legert et al., 2014). Despite the previously published risks of impaired wound healing in patients treated with sirolimus, we found no differences in the course of OM after HSCT.
The courses of healing in those receiving Tac/Sir and those receiving CsA/Mtx were similar; the OMAS score had returned to baseline after 24 days. This is in contrast to Cutler et al. (2014), who found more rapid healing of oropharyngeal mucositis in the Tac/Sir group.
Wound-healing complications associated with sirolimus therapy are typically diagnosed within the first few months following transplant (Troppmann et al., 2003). In the kidney transplant setting, because of low nephrotoxicity, the use of sirolimus in combination with a calcineurin inhibitor has resulted in a low incidence of graft rejection, a lowered occurrence of new malignancies and a possibility of steroid withdrawal among high immune responder kidney transplant recipients (Kahan, 2000). At the same time, the use of sirolimus has been associated with increased wound-healing complications, defined as lymphocele, wound infection and incisional hernia (Knight et al., 2007). The 2014 study (Garming Legert et al., 2014) was done on a larger cohort than in the present study and found, in addition to these two risk factors, that female-to-male transplantation and being seropositive for 3-4 different herpes viruses were also risk factors.
The significantly increased risk of mucositis in patients treated with MAC as opposed to RIC is in line with a previous prospective randomized study in HSCT patients with myeloid malignancies (Ringdén et al., 2013). Patients randomized to MAC in the present study had a significantly higher median mucositis grade (OM 4) than RIC patients (OM 1). Several other toxicities were also increased in the MAC group. The strengths of this study include an RCT study design and randomization of the participants. Other strengths include the large size of the cohort of patients who underwent HSCT and the more than 3 weeks of continuous follow-up with oral examinations. Limitations include a lack of stratification of patients during randomization concerning level of HLA match, which is a known risk factor of OM in HSCT settings. No blinding was attempted after randomization. OM was analysed using the WHO toxicity scale, although the OMAS scale is more often used in research. The two scales, however, have good correlation.
In conclusion, the present study found that, as a GVHD prophylaxis, tacrolimus/sirolimus following HSCT increased neither the incidence nor severity of OM compared with cyclosporine/methotrexate. Furthermore, we show that OM healing was not prolonged and followed the same time pattern as cyclosporine/methotrexate.