Dysphagia Symptom Score for Eosinophilic Esophagitis
Dysphagia Symptom Score for Eosinophilic Esophagitis
Dysphagia is a hallmark of EoE in adolescents and adults. As such, it is an attractive symptom outcome for treatment trials, but no validated dysphagia instruments exist for EoE. The aim of this study was to develop and field test a PRO for dysphagia in adolescents and young adults with EoE that would meet FDA developmental guidelines and could be used as a reliable outcome measure. The result of this study, the DSQ, is a daily symptom diary that was developed with a combination of qualitative and quantitative research techniques. The DSQ accurately assesses dysphagia frequency and severity in one dimension, and has content validity. It also showed differences between adolescents and adults, and between those treated and not treated with topical corticosteroids.
There have been a number of previous efforts to accurately measure symptoms in EoE. In the first randomised control trial of fluticasone compared with placebo in children with EoE, Konikoff et al. assessed a wide variety of symptoms associated with EoE, including nausea, vomiting, abdominal pain, chest pain, heartburn, regurgitation, dysphagia, food impaction, poor appetite and poor weight gain. Symptoms were defined as either present or absent and a scoring system was not used. Aceves et al. proposed a symptom scoring tool for children with EoE that incorporated multiple symptoms, their frequency and their severity. They found that symptoms of dysphagia and early satiety correlated with endoscopic and histological findings, but that other symptoms did not. A version of this score was used in a randomised controlled trial (RCT) of budesonide compared with placebo to treat children with EoE, and subjects treated with budesonide had improvement in symptoms, whereas those in the placebo group did not.
Although these studies in children assessed multiple symptoms, several studies in adults have focused solely on dysphagia. In an RCT conducted by Straumann et al. which compared budesonide with placebo for treatment of EoE in adults, the SDI was used to measure the frequency and intensity of dysphagia. Those treated with budesonide had an improvement in symptoms related to dysphagia that was not seen in the placebo group. Several other studies of fluticasone and budesonide for treatment of EoE in adults have used the Mayo Dysphagia Questionnaire-30 Day (MDQ-30) to assess dysphagia. Although the MDQ-30 was developed and validated for peptic strictures, none of the previously used measures have been validated for use in EoE, and as far as we are aware, none went through a development phase in a manner consistent with the FDA guidance on PROs.
More recently, Franciosi et al. have published a series of papers on the development and validation of the Pediatric Eosinophilic Esophagitis Symptom Score and associated quality of life measures. Although these instruments are being developed in concert with FDA collaboration, they are not yet available for use in EoE treatment trials for children, and would not be applicable to adults. Quality of life measures informed by formative methods, as well as a multi-domain symptom activity index, are also being developed and validated in adults, but are similarly not yet available for use.
The rationale for carefully developing a well-defined and reliable PRO instrument is directly related to interpreting the validity of results from clinical trials. If a pharmacological agent does not appear to provide a symptom benefit, but the symptom was measured with an unreliable instrument, it is difficult to know whether the treatment itself was not effective or if the outcome was not accurately measured. This has been a major challenge in EoE trial design and interpretation. Randomised studies of budesonide, fluticasone, anti-IL-5 antibodies and a novel antagonist to the chemoattractant receptor-homologous molecule expressed on Th2 cells have all shown histological benefit in decreasing oesophageal eosinophil counts. However, in these studies, histological response was not correlated with symptom response. Although there are several potential reasons for this, including symptom/histology discordance being an intrinsic feature of EoE, misclassification of non-EoE patients as EoE cases, and that the understanding of EoE is still in the early stages, another possible explanation is that the nonvalidated symptom measures did not perform as intended.
It is important to acknowledge that this study has some limitations. First, it was a development study, and although content validity and item reliability were confirmed, there is still a need to determine responsiveness to therapy and reproducibility in larger treatment trials. Second, the study had a relatively small sample size with prevalent EoE cases who were on a variety of treatments. However, for the purposes of concept elicitation, cognitive interviews and field testing the DSQ instrument, a homogenous population of EoE cases is not required or even desired as a mix of disease severity helps to provide broader context for the development of the questionnaire. Third, the racial characteristics of the study population (nearly all White) may not reflect the general population of EoE cases, particularly at centres in racially diverse areas. Fourth, the DSQ does not assess for dietary modification. Although we had hypothesised that this would be potentially an important behaviour to ask about, it did not emerge as a consistent theme among the patients during the first phase of the study, and therefore a question related to dietary modification was not included in the final DSQ. Fifth, the sub-group comparison that showed that the DSQ distinguished current topical steroid users from those not currently using topical steroids should be interpreted with caution. These subjects were not randomised to a treatment, and although the assumption could be made that those on topical steroids had less dysphagia due to the medication, there could be other reasons such as bias or confounding that could explain this finding. Although the current steroid users were younger, after accounting for age, the trend for less dysphagia in steroid users still held.
There are also a number of strengths to this study. It was a multi-centre prospective study that was explicitly designed to meet the FDA guidance for development of PROs. The DSQ uses a short (daily) recall window to minimise recall bias and maximise reporting of the most current symptoms, and the electronic format maximises data capture and accuracy. The age range of subjects allowed for self-reporting using the patient's language. The focus on the single, predominant symptom of dysphagia in a defined patient age group is the most likely strategy to result in a responsive instrument. In addition, the DSQ correlated strongly with the SDI, a dysphagia score that was responsive in a clinical trial.
In conclusion, this study successfully developed the DSQ, a three-question PRO in an electronic format, in accordance with the FDA guidelines. This instrument measures dysphagia frequency and intensity in adolescents and adults with EoE in a single domain with a summary score. The DSQ has face and content validity, has been judged easy to use with excellent compliance and is ready to be used in clinical trials as a clinical outcome measure.
Discussion
Dysphagia is a hallmark of EoE in adolescents and adults. As such, it is an attractive symptom outcome for treatment trials, but no validated dysphagia instruments exist for EoE. The aim of this study was to develop and field test a PRO for dysphagia in adolescents and young adults with EoE that would meet FDA developmental guidelines and could be used as a reliable outcome measure. The result of this study, the DSQ, is a daily symptom diary that was developed with a combination of qualitative and quantitative research techniques. The DSQ accurately assesses dysphagia frequency and severity in one dimension, and has content validity. It also showed differences between adolescents and adults, and between those treated and not treated with topical corticosteroids.
There have been a number of previous efforts to accurately measure symptoms in EoE. In the first randomised control trial of fluticasone compared with placebo in children with EoE, Konikoff et al. assessed a wide variety of symptoms associated with EoE, including nausea, vomiting, abdominal pain, chest pain, heartburn, regurgitation, dysphagia, food impaction, poor appetite and poor weight gain. Symptoms were defined as either present or absent and a scoring system was not used. Aceves et al. proposed a symptom scoring tool for children with EoE that incorporated multiple symptoms, their frequency and their severity. They found that symptoms of dysphagia and early satiety correlated with endoscopic and histological findings, but that other symptoms did not. A version of this score was used in a randomised controlled trial (RCT) of budesonide compared with placebo to treat children with EoE, and subjects treated with budesonide had improvement in symptoms, whereas those in the placebo group did not.
Although these studies in children assessed multiple symptoms, several studies in adults have focused solely on dysphagia. In an RCT conducted by Straumann et al. which compared budesonide with placebo for treatment of EoE in adults, the SDI was used to measure the frequency and intensity of dysphagia. Those treated with budesonide had an improvement in symptoms related to dysphagia that was not seen in the placebo group. Several other studies of fluticasone and budesonide for treatment of EoE in adults have used the Mayo Dysphagia Questionnaire-30 Day (MDQ-30) to assess dysphagia. Although the MDQ-30 was developed and validated for peptic strictures, none of the previously used measures have been validated for use in EoE, and as far as we are aware, none went through a development phase in a manner consistent with the FDA guidance on PROs.
More recently, Franciosi et al. have published a series of papers on the development and validation of the Pediatric Eosinophilic Esophagitis Symptom Score and associated quality of life measures. Although these instruments are being developed in concert with FDA collaboration, they are not yet available for use in EoE treatment trials for children, and would not be applicable to adults. Quality of life measures informed by formative methods, as well as a multi-domain symptom activity index, are also being developed and validated in adults, but are similarly not yet available for use.
The rationale for carefully developing a well-defined and reliable PRO instrument is directly related to interpreting the validity of results from clinical trials. If a pharmacological agent does not appear to provide a symptom benefit, but the symptom was measured with an unreliable instrument, it is difficult to know whether the treatment itself was not effective or if the outcome was not accurately measured. This has been a major challenge in EoE trial design and interpretation. Randomised studies of budesonide, fluticasone, anti-IL-5 antibodies and a novel antagonist to the chemoattractant receptor-homologous molecule expressed on Th2 cells have all shown histological benefit in decreasing oesophageal eosinophil counts. However, in these studies, histological response was not correlated with symptom response. Although there are several potential reasons for this, including symptom/histology discordance being an intrinsic feature of EoE, misclassification of non-EoE patients as EoE cases, and that the understanding of EoE is still in the early stages, another possible explanation is that the nonvalidated symptom measures did not perform as intended.
It is important to acknowledge that this study has some limitations. First, it was a development study, and although content validity and item reliability were confirmed, there is still a need to determine responsiveness to therapy and reproducibility in larger treatment trials. Second, the study had a relatively small sample size with prevalent EoE cases who were on a variety of treatments. However, for the purposes of concept elicitation, cognitive interviews and field testing the DSQ instrument, a homogenous population of EoE cases is not required or even desired as a mix of disease severity helps to provide broader context for the development of the questionnaire. Third, the racial characteristics of the study population (nearly all White) may not reflect the general population of EoE cases, particularly at centres in racially diverse areas. Fourth, the DSQ does not assess for dietary modification. Although we had hypothesised that this would be potentially an important behaviour to ask about, it did not emerge as a consistent theme among the patients during the first phase of the study, and therefore a question related to dietary modification was not included in the final DSQ. Fifth, the sub-group comparison that showed that the DSQ distinguished current topical steroid users from those not currently using topical steroids should be interpreted with caution. These subjects were not randomised to a treatment, and although the assumption could be made that those on topical steroids had less dysphagia due to the medication, there could be other reasons such as bias or confounding that could explain this finding. Although the current steroid users were younger, after accounting for age, the trend for less dysphagia in steroid users still held.
There are also a number of strengths to this study. It was a multi-centre prospective study that was explicitly designed to meet the FDA guidance for development of PROs. The DSQ uses a short (daily) recall window to minimise recall bias and maximise reporting of the most current symptoms, and the electronic format maximises data capture and accuracy. The age range of subjects allowed for self-reporting using the patient's language. The focus on the single, predominant symptom of dysphagia in a defined patient age group is the most likely strategy to result in a responsive instrument. In addition, the DSQ correlated strongly with the SDI, a dysphagia score that was responsive in a clinical trial.
In conclusion, this study successfully developed the DSQ, a three-question PRO in an electronic format, in accordance with the FDA guidelines. This instrument measures dysphagia frequency and intensity in adolescents and adults with EoE in a single domain with a summary score. The DSQ has face and content validity, has been judged easy to use with excellent compliance and is ready to be used in clinical trials as a clinical outcome measure.