Prevalence and risk factors of oral mucositis in paediatric patients undergoing haematopoietic stem cell transplantation

Abstract Background A complete understanding of oral mucositis (OM) is crucial to develop appropriate interventions to aid in the successful overall health outcome of paediatric patients undergoing haematopoietic stem cell transplantation (HSCT). Aims This study aimed at determining the prevalence and severity of OM and at identifying the predictive factors that might aggravate OM at one‐week, two‐week and three‐week post‐HSCT. Methods This retrospective, hospital‐based study reviewed the medical records of 170 paediatric patients, summarising the patients’ characteristics using descriptive statistics. Binary logistic regression was used to identify factors associated with the development of OM. Results At one‐week post‐HSCT, 41% of 140 patients (n = 49) had developed OM, this was reduced at two‐week (n = 36, 33%) and three‐week (n = 13, 19%) post‐HSCT. Univariate logistic regression revealed that patients with cancer (OR = 0.16, 95% CI = 0.05–0.54; p‐value = .003) had a significantly lower prevalence of OM. Younger patients with an average age of 7.9 years old (OR = 0.85, 95% CI = 0.75–0.97; p‐value = 0.013) and the presence of GvHD (OR = 2.37, 95% CI = 1.03–5.45, p‐value = 0.042) were significantly related to a higher prevalence of OM. Multivariable logistic regression confirmed that the risk of OM is lower in patients with cancer compared to those with immunodeficiency syndromes or hereditary blood diseases (OR = 0.18, 95% CI = 0.04–0.77; p‐value = .021). Conclusions This study identified a significantly lower prevalence of OM in patients with cancer compared to other conditions and that young recipients and those who developed GvHD were more likely to have OM.

. Haematopoietic stem cell transplantation (HSCT) is the treatment of choice for various malignancies, immune deficiencies and bone marrow failure syndromes in children, adolescents and adults (Barriga et al., 2012;Copelan, 2006). OM is arguably the single most painful and debilitating complication in patients undergoing high-dose chemotherapy preparatory to HSCT (Bellm et al., 2000), with the incidence of OM among patients receiving both autologous and allogeneic HSCT varying between 75% and 99%, though OM tends to be more prevalent among paediatric patients (Eduardo et al., 2015;Vagliano et al., 2011).
OM is a predisposing factor for the development of further adverse implications, including an increased risk for local and systemic infections, poor nutrition and prolonged hospitalisation, as well as periodontal diseases, dental caries, dry mouth, trismus, dysphagia and dysgeusia (Mubaraki, 2019). Consequently, OM has a marked impact on the quality of life and cost of treatment (Mubaraki, 2019).
The most invalidating complications of OM are encountered during the conditioning regimen prior to HSCT (Cinausero et al., 2017), and depending on the severity, OM might necessitate the interruption and/or dose reduction of anticancer therapy (Berger et al., 2018).
However, little is known about the relevant factors associated with increased susceptibility of individuals to severe OM, specifically following HSCT (Bowen & Wardill, 2017). Complete healing of OM lesions typically occurs without scar formation unless it is exacerbated by a severe infection (Köstler et al., 2001).
Pathogenesis of OM is characterised by a complex five-step biological process beginning with initiation, followed by primary damage response, signal amplification and eventually ulceration and healing (Sonis, 2004). OM is influenced by several risk factors which are usually classified into two main categories: patient-related (i.e. age and type of malignancy) and treatment-related (i.e. total body irradiation and type of cytotoxic agent) (Barasch & Peterson, 2003).
Nevertheless, some prophylactic measures have been identified to minimise the intensity of OM (i.e. keratinocyte growth factor and cryotherapy) with currently no universally validated methods for prophylactic or therapeutic measures in children (Bowen & Wardill, 2017;Worthington et al., 2011). Also, the administration of opioid analgesics is often required to manage moderate to severe pain in children (Kuiken et al., 2015). Therefore, a complete understanding of this condition is crucial to develop appropriate interventions to aid in the successful overall health outcome. Given the scarce data available regarding OM in paediatric patients undergoing HSCT, this retrospective, hospital-based study aimed at determining the prevalence and severity of OM, and at identifying the predictive factors that might aggravate OM at one-week, two-week and three-week post-HSCT.

| Ethical approval
Ethical approval was granted from the Institutional Review Board of the King Faisal Specialist Hospital and Research Centre (KFSH&RC), Riyadh, Saudi Arabia (RAC number: 2091015) before commencing any study-related procedures.

| Study design and setting
This was a retrospective analysis of the medical records of paediatric patients treated with HSCT at the Department of Haematology-Oncology and Stem Cell Transplantation, KFSH&RC, Riyadh, Saudi Arabia.

| Inclusion and exclusion criteria
Paediatric patients (0 to ≤ 14 years old) who received either autologous or allogeneic HSCT for the treatment of cancer (i.e. acute lymphocytic leukaemia, acute myelocytic leukaemia and solid tumours), hereditary blood disease (i.e. sickle cell anaemia, Fanconi anaemia and thalassaemia) and immune deficiency syndromes were eligible for inclusion in this study. Patients who had an oral health assessment for OM post-HSCT were included in this study. Patients above 14 years of age and those who did not receive HSCT as part of the treatment or did not have oral health assessment for OM post-HSCT were excluded from the study.

| Procedure
The medical records of all patients (n = 170) who received HSCT in the study period were reviewed by the direct care team from the Dental Department at KFSH&RC, Riyadh, Saudi Arabia. As part of standard care in this study centre, all patients undergoing HSCT received instructions to use supersaturated calcium phosphate rinse and an extra-soft toothbrush twice a day with their existing oral hygiene protocol regimen (0.2% chlorhexidine gluconate + 3% sodium bicarbonate + nystatin 100,000 U/ml). All patients received a unique ID (i.e. numbers from 1 to 170); thus, the patient's identity remained anonymous throughout the study, as the code number was only known to the direct care team. Patients' demographic details (i.e. age and gender) and clinical parameters (i.e. medical condition, type of HSCT (i.e. allogeneic or autologous), presence of the graft versus host diseases (GvHD), presence and grade of OM and the average time between conditioning regimen administration and HSCT) and haematological parameters (i.e. white blood cells (WBCs), haemoglobin (HB) level and platelet count) were extracted from the patients' medical records. All data were transferred into a Microsoft Excel spreadsheet. In this study, demographic and clinical/haematological parameters were recorded at three different times (i.e. one-week, two-week and three-week post-HSCT.

| Measures
A limited set of well-trained staff nurses was responsible for the assessment and documentation of OM, which was graded using the World Health Organization (WHO) oral toxicity fifth-grade scale (WHO, 1979). The severity of OM was graded as follows: Grade 0 = solid diet, no soreness, no erythema, no ulcers, Grade I = solid diet, soreness and/or erythema, Grade II = solid diet, erythema and/ or ulcers, Grade III = liquid diet, erythema and/or ulcers, and lastly Grade IV = unable to swallow, erythema and/or ulcers. Utilisation of this scale ensured ease of use, validity and reliability.

| Statistical analysis
A descriptive analysis was performed to define the characteristics of the study sample through a form of counts and percentages. The normality of the data was checked using a histogram, Kolmogorov-Smirnov Lillefors and Levene's tests. A chi-square test was used to establish a relationship between categorical variables. Accordingly, haematological parameters were compared between males and females using a parametric test (i.e. independent t tests), while the nonparametric Kruskal-Wallis test was used to compare haematological parameters and medical conditions (i.e. cancer, hereditary blood diseases and immune deficiency syndromes). Binary logistic regression analysis was applied to identify factors associated with the occurrence of OM at univariate and multivariable levels. The statistical significance was assumed at a 5% level, and the statistical analysis was performed using the Statistical Package for Social Science (Released 2015, IBM SPSS Statistical for Windows, Version 23.0, Armonk, NY: IBM Corp).

| Sample size
A minimum sample of 143 was enough to assess whether there is a significant difference in the prevalence of OM at three different times (i.e. one-week, two-week and three-week post-HSCT). The sample size was calculated based on X 2 tests (Goodness-of-fit tests: Contingency tables) at the 5% level of significance (α err prob = 0.05) with 80% power (1β err prob = 0.80), medium effect size (effect size w = 0.30), and Df of 5. The G*power 3.1.9.2 was used to calculate the sample size.

| Sample characteristics
Of the 170 individuals, data of 140 eligible patients were included in the study analysis. An overview of their demographic characteristics is provided in Table 1. The average age was 8.8 years old (SD = 3.47), and most patients (n = 76, 54%) were male. In terms of medical conditions, some had been diagnosed with cancer (n = 35, 25%), and others had been diagnosed with hereditary blood diseases (n = 32, 23%), with most (n = 73, 52%) diagnosed with immunodeficiency syndromes. The vast majority of patients had allogeneic HSCT (n = 134, 96%) and did not develop GvHD (n = 93, 66%), with an average time between conditioning regimen administration and HSCT of 7.0 days (SD = 2.88). Table 2 compares the haematological parameters of patients according to gender. Independent t test showed that there were no significant differences between males and females regarding the WBCs, HB level and platelets count at one-week, two-week and three-week post-HSCT (p-value > .05).

| Multivariable logistic regression analyses
Multivariable logistic regression confirmed that the risk of OM decreases in patients with cancer compared to those with immunodeficiency syndromes or hereditary blood diseases independently (OR = 0.18, 95% CI = 0.04-0.77; p-value = .021)-see Table 7.
Using direct entry, the Nagelkreke R 2 was 0.21, indicating that the

| D ISCUSS I ON
OM in paediatric patients constitutes a major oncological dilemma with practical limitations for assessment tools and treatment methods (Farrington & Cullen, 2010). Based on the recommendation of the literature to further investigate the risk prediction of OM, especially among the paediatric population (Bowen & Wardill, 2017), the current retrospective study aimed at determining the prevalence and severity of OM and at identifying the predictive factors that might aggravate OM early after HSCT infusion.
A recently published study pointed out that the incidence of OM among children (4-17 years) after HSCT (all grades of severity) was 80% (Kamsvåg et al., 2020). Similarly, a multi-centre study of 262 children/adolescents aged between 0 to 18 years old reported that 79.8% developed OM (Vagliano et al., 2011). In contrast, the current study found that the prevalence of OM among paediatric patients (≤14 years) was 41%, making our findings inconsistent with those in the literature. This contradictory result might be due to the inability to assess the oral cavity properly in uncooperative and extremely young patients, as well as the incapacity of the child to attribute pain to bodily locations, or rather because of the prevention protocols that were implemented in the study centre, which we believe may have influenced the onset of OM, as those oral hygiene regimens were suggested to be of benefit in the reduction of OM occurrence The present study findings indicated that younger patients are significantly more prone to have OM at the two-week post-HSCT, similar to a previous report, wherein OM was found to be significantly influenced by the recipient age (Bardellini et al., 2013). A possible explanation for this might be linked to the high proliferative rate of basal epithelial cells in younger individuals (Sonis, 2007(Sonis, , 2011. When the three underlying categories of medical conditions were compared, the prevalence of OM in the second-week post-HSCT was significantly lower in patients with cancer, whereas immunocompromised patients showed the greatest prevalence, which may be related to the type of conditioning regimen used rather than to the disease itself. For instance, a prospective evaluation study indicated that the principal determinant of OM was the preparative conditioning regimen (Wardley et al., 2000).
Therefore, a powered, prospective study that focuses on the effect of the conditioning regimen with regard to the type of medical condition on the risk of OM development following HSCT is needed.
In our study, the presence of GvHD was significantly related to the development of OM. However, a retrospective study found no association between GvHD and OM, as they reported that the use of prophylaxis against GvHD was more likely to induce OM rather than GvHD itself (Bardellini et al., 2013). It is worth mentioning that Bardellini and his colleagues used only descriptive statistics;  Ref.

| LIMITATI ON S AND FUTURE RE S E ARCH
One of the limitations of the current study was the retrospective nature of the data analysis, which could have contributed to missing data.
Additionally, this was a single-centre study; hence this limits the generalisability of the results. Also, some potential confounding factors were not addressed in this study, which could influence OM development such as conditioning regimen, GvHD prophylaxis, nutritional route and whether or not the patients were hospitalised. Future studies may incorporate multiple centres to address the above-mentioned variables to overcome the current limitations.

| CON CLUS I ON S AND RECOMMENDATIONS
Our study identified that young recipients and those who developed GvHD were more likely to have OM post-HSCT. Moreover, the findings confirmed a significant decrease in the prevalence of OM in patients with cancer, raising the possibility that cancer patients might have obtained superior benefits of HSCT treatment relevant to their oral health compared to patients with other medical conditions who deserve equal or even greater attention. Therefore, comprehensive pre-and post-HSCT oral care should also be emphasised for young patients with immunodeficiency syndromes or hereditary blood diseases. There is also a need for continued efforts directed to researching the risk prediction among this group of patients to develop intervention strategies and to establish evidence-based guidelines for proper management of this condition.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/odi.13777.