Novel Simplified Nasal Endoscopy Grading System for Hereditary Hemorrhagic Telangiectasia Patients

Authors: Ethan Soudry, Nimrod Amitai, Aiman Elmograbi, Igor Vainer, Meir Mei‐Zahav

Abstract

Objective

A validated and practical grading system for endoscopic findings is crucial for both clinical care and research when evaluating patients with Hereditary Hemorrhagic Telangiectasia (HHT) who experience epistaxis and nasal complaints. This study aimed to validate a novel, simplified nasal endoscopy grading system in HHT patients.

Methods

The study was conducted at a tertiary referral center for HHT patients. Twenty HHT patients were derived from an earlier double‐blinded, placebo‐controlled study by our group, investigating the efficacy of topical propranolol gel for treating epistaxis.

Patients were examined at three time points, and nasal endoscopies were recorded. The grading system categorized findings into three mild (few punctate lesions), moderate (multiple telangiectasias/large arteriovenous malformations involving the anterior nasal septum), and severe (diffuse involvement of the nasal mucosa with telangiectasias).

Three experienced rhinologists, blinded to clinical data, performed the grading. Primary outcomes were the correlation of the grading system with Epistaxis Severity Score (ESS) and Quality of Life (QoL). Inter‐rater reliability was assessed using the intraclass correlation coefficient (ICC).

Results

54 nasal endoscopy scores from 20 patients were analyzed. There was a significant correlation between the grading system and ESS (r = 0.8, p = 0.006) and QoL (r = −0.42, p = 0.002). Worse ESS and QoL scores were associated with more severe nasal endoscopy grades. The ICC of 0.8 (p < 0.0001) indicated substantial agreement among the three graders.

Conclusions

The novel nasal endoscopy grading system demonstrated a significant correlation with ESS and QoL. Further validation of this novel grading system in larger cohorts is needed.

Level of Evidence

3

Introduction

Hereditary hemorrhagic telangiectasia (HHT), formerly known as Osler–Weber–Rendu syndrome, is a rare autosomal dominant disorder with variable penetrance, characterized by vascular dysplasia. This causes mucocutaneous telangiectases and visceral arteriovenous malformations (AVMs) [1, 2]. Almost all patients (95%) suffer from recurrent epistaxis of varying degrees of severity, ranging from occasional bleeding to life‐threatening blood loss [1, 2, 3]. Mean age of onset of epistaxis in HHT patients is 12 years and represents the most common cause of disability and social impairment; an association between its severity and a lower quality of life (QoL) has been found [4, 5].

Although epistaxis is considered the cardinal sign of HHT and the most common cause of decreasing QoL in these patients, only a few studies have attempted to develop tools to facilitate staging of telangiectases involving the nasal mucosa [6, 7, 8, 9]. Reh et al. [6] formulated a novel endoscopic staging system (HHT endoscopic score – HES) comprising 7 different parameters. While the HES was shown to strongly associate with the epistaxis severity score (ESS) it has not gained wide acceptance, likely since it is cumbersome and impractical in the clinical setting. Mahoney et al. [7], proposed a three‐staged classification of intranasal vascular patterns in HHT patients as a means for selection of Nd‐YAG laser treatment, but have not evaluated its association with health‐related quality of life. Pagella et al. [9] observed that telangiectasias endoscopically appearing as large and prominent correlate with a higher frequency of epistaxis. Haubner et al. [8] observed that berry‐like telangiectasias are significantly associated with lower quality of life.

Our rationale for developing a new grading system is based on the need for a more simplified, practical, reproducible, and reliable widely applicable endoscopic scoring tool that directly correlates with clinical measures of disease severity, such as ESS and QoL. While the Mahoney classification provides a useful framework for treatment stratification, it does not comprehensively evaluate the burden of disease from a patient‐centered perspective. In addition, the existing grading system of Reh et al. is too complex for everyday practice. Thus, there is a need for a new and simplified grading system that can be easily adopted in routine follow‐ups and used to guide management decisions.

In our study, we aimed to validate a novel simplified nasal endoscopic grading system in HHT patients in correlation to clinical parameters such as the ESS and QoL score.

Methods

Patients and Clinical Data

This observational cross‐sectional study was derived from an earlier double‐blinded placebo‐controlled study conducted by our group, investigating the efficacy of topical propranolol gel in the treatment of epistaxis in HHT patients [10]. The original study was comprised of a double‐blinded placebo‐controlled phase, with a subsequent open‐label phase. Patients previously diagnosed with HHT were recruited from our National HHT center. Study exclusion and inclusion criteria are summarized in Table 1. In brief, the inclusion criteria for the parent randomized controlled trial required moderate‐to‐severe epistaxis (ESS > 4). This ensured that our cohort included patients with substantial disease burden. In practice, all HHT patients who presented during the study period were enrolled as long as they met the inclusion criteria, minimizing selection bias.

The study was initiated after approval by the local internal review board and Ethics committee (RMC‐0191‐15). All participants signed an informed consent form.

Following screening, the patients were randomly assigned, at a 1 ratio, to be treated with propranolol gel or a placebo gel for a period of 8 weeks (the double‐blind phase). Following the double‐blind phase, all the patients were offered to continue an open‐label phase for an additional 8 weeks.

The study participants were examined at screening and randomization and at the end of each treatment phase (the 8th and 16th weeks). Each visit included a recent medical history and physical examination; monitoring of side effects; measurements of heart rate (HR), blood pressure (BP), HB, and iron and ferritin levels; a rhinology examination; assessment of epistaxis severity according to the epistaxis severity score (ESS); and health‐related QoL according to the 12‐Item Short Form Health Survey (SF‐12) questionnaire. An electrocardiogram was performed at screening and repeated at the primary investigator's discretion.

Nasal endoscopies were recorded at screening and at the 8th and 16th weeks, and representative photographs of the nasal cavities were captured in a de‐identified manner. Patients' nasal cavities were decongested with lidocaine 1.5% and phenylephrine 1% spray prior to endoscopic examination with a zero‐degree 4 mm endoscope connected to a high‐definition camera and monitor (Storz). This decongestion step was performed to improve patient comfort and visualization of the nasal mucosa. When dried blood clots were noted to obscure the endoscopy and prevent accurate grading, gentle removal of these clots was undertaken.

Nasal involvement with disease at recruitment was graded as mild–few punctate telangiectases involving the nasal cavity (grade I, Figure 1a), moderate–multiple telangiectases involving the nasal cavity/large arteriovenous malformations involving the anterior nasal septum (grade II, Figure 1b–), and severe‐diffuse involvement of the nasal mucosa with telangiectases (grade III, Figure 1c). Notably, telangiectases corresponding to Grades I and II may involve any area in the nasal cavity (although more common on the anterior nasal septum). When grading was not similar between sides, grading was determined according to the worst side.

All patient images were subsequently graded at the conclusion of the original RCT [10] by the first author (ES) of this study who conceptualized the grading score. In order to evaluate the reliability of this new grading system, two additional expert rhinologists (NA, AEM) not involved in the original RCT were asked, long after the completion of the RCT, to participate in this current study and grade in an anonymized fashion the exact same patient endoscopic images.

The primary outcome was a statistically significant correlation of our novel endoscopic grading system with quality of life measures and clinical findings.

Reliability and Validation

Inter‐rater reliability between three graders was calculated for all examinations.

Statistical Analysis

Data was analyzed using the SPSS statistical software version 25.0 (SPSS Inc., Cary, NC, USA). For the analysis of continuous data, and due to the sample size, Kruskal–Wallis was applied. Chi‐Square or Fisher's test were utilized for the analysis of categorical variables. When performing correlation analysis, Spearman's correlation was chosen due to the nature of the data. Inter‐rater reliability was assessed using intraclass correlation coefficient (ICC) analysis.

Finally, a linear regression analysis, adjusted for variables that reached a significance level of 0.05 in univariate analysis, was included in order to decrease confounding variables effect. A two‐sided p value < 0.05 was considered statistically significant.

Results

Twenty participants completed the double‐blind phase, with 10 participants in each study group. At presentation, the mean age of the cohort was 54.05 ± 10.72 years, and 75% were women. The baseline ESS score was (6.03 ± 1.72) and the baseline QOL score was (34.93 ± 8.84); additional baseline characteristics are summarized in Table 2.

The reliability of the grading system across the three observers demonstrated an intraclass correlation coefficient (ICC) of 0.8 (p < 0.0001), indicating substantial agreement. A total of 54 nasal endoscopy scores were available for analysis. Significant correlations were observed across the dataset, with endoscopic grades positively correlating with ESS (r = 0.41, p = 0.001), and inverse correlations were noted between endoscopic grades and QOL (r = −0.43, p < 0.001). No significant correlation was observed with Hemoglobin (Hb) levels, as observed in Table 3. Significantly worse ESS and QoL scores were observed incrementally with a more severe nasal endoscopy grade, as depicted in Table 4.

Further subgroup analysis of the study's treatment arms, as shown in Table 5, revealed that in the control group (no gel, n = 20 endoscopies), endoscopic grades were significantly correlated with ESS (r = 0.52, p = 0.02) and QOL (r = −0.57, p = 0.01).

In the treatment group (propranolol gel, n = 34 endoscopies), endoscopic grades remained significantly correlated with ESS (r = 0.41, p = 0.02), while the correlation with QOL was not statistically significant (r = −0.32, p = 0.08).

To address confounding variables and adjust our grading system for these parameters, we used simple linear regression, adjusting ESS for the new endoscopic grade and QOL. After accounting for these factors in the model, the new grading system continued to show a significant correlation to the ESS (r = 0.8, p = 0.006).

Discussion

Development of a novel, easy‐to‐use nasal endoscopic grading system for HHT patients, which can be easily replicated in and used in multiple clinical settings, is of paramount importance for the decision‐making process when considering treatment options and for consistency and common terminology in research reports. In this study, we found our novel endoscopic grading system to be significantly correlated with ESS score (r = 0.8, p = 0.006) with significant correlation between the study's three graders. Moreover, the simplicity of our grading system, being based on a three‐grade staging, allows ease of utilization and supports using it in the clinical setting during initial intake as well as during follow‐up.

Interestingly, very few studies reported classification systems for intranasal vascular patterns in HHT patients. The first one, by Mahoney et al. [7], suggested a simplified grading system as a means for treatment selection–Nd‐YAG laser for photocoagulation treatment or septodermoplasty. Their classification stratified intranasal lesions to (I) isolated punctate telangiectasias or individual small AVM; (II) diffuse interconnecting vasculature with “feeder”; (III) large solitary AVM, which may be associated with scattered telangiectasia. In that study, the authors found that patients with patterns I and III showed greater improvement in epistaxis after Nd‐YAG laser photocoagulation, and patients with pattern II fared better with septodermoplasty. Nevertheless, that study did not assess its grading classification with correlation to functional outcomes such as epistaxis severity, hemoglobin levels, or QoL.

Pagella et al. [9], using two graders, proposed a grading system that defined three types of nasal “punctate pattern,” when all subsites involved present only flat punctate telangiectasias; “large pattern,” when all subsites involved present only prominent large lesions; and the “mixed pattern,” when both flat punctate and prominent large telangiectasias are present in the nasal fossae. They observed that telangiectasias endoscopically appearing as large and prominent correlate with a higher frequency of epistaxis. An increase in the number of nasal subsites involved correlated with a higher intensity of nosebleeds. As the ESS score was not available at the time of the publication of this work, it is not known whether there is a correlation between this grading system and the ESS score.

Reh et al. [6], using a single grader, reported the HHT endoscopy score (HES score) and found it to be strongly correlated with the ESS score. This scoring system, however, includes four stages, and each stage is determined according to seven number of sites involved, vascular morphology patterns, presence of crusting, location of vascular lesions, presence of perforation, density of lesions, and presence of relative AVMs. This classification, however, has not gained wide acceptance as a useful tool in both clinical practice and research reports, probably due to its complexity.

Haubner et al. [8], using three graders, proposed a new classification system for endonasal HHT lesions using digital microscopy, demonstrating that raised, berry‐like telangiectasias are significantly associated with lower quality of life using VAS score. However, no statistically significant association was noted with ESS score. Moreover, digital microscopy is not the standard tool used in everyday clinical practice as compared to nasal endoscopy.

Our novel grading system emphasizes ease‐of‐use and significant correlation with clinically significant factors (ESS and QoL). Additionally, it is the first grading system for HHT severity, which was tested for reproducibility by three experienced rhinologists after formulation, with good agreement (ICC = 0.8), which strengthens its reliability. Thus, its simplicity and reproducibility may facilitate better communication and decision‐making in relation to HHT patients.

Interestingly, our grading system was not significantly correlated with Hb levels. Nonetheless, Hb levels may be affected by multiple factors such as diet, comorbidities, and bleeding from other sources, such as the gastrointestinal system. Similar results were also observed in the study by Haubner [8]. Of note, the other previously published works assessing grading systems had not assessed correlations between their proposed grading systems and hemoglobin levels.

Although this study included both control and propranolol gel treatment groups, significant correlations remained between endoscopic grading score and ESS scores in both groups thus further underscoring the relevance of this novel endoscopy score. Nevertheless, a significant correlation between the endoscopic grading system and QoL score was observed only in the control group.

There are several limitations in this study worth noting. One limitation stems from the rarity of HHT in the general population with a prevalence that ranges between 5000 and 10,000. Due to the low prevalence of HHT, the study population was relatively small. Another limitation is due to the design of the original study [11], which included the application of propranolol gel in the nose, which might have affected the morphology of the nasal lesions and introduced bias. The original study design may have also introduced selection bias and a placebo effect for ESS and QoL.

Conclusion

Recurrent epistaxis is the most prevalent presentation of HHT and has an enormous impact on patient QoL. This study proposes a simple and highly practical novel endoscopic grading system of the severity of nasal lesions in HHT patients that highly correlates with the ESS score and which can greatly facilitate communication between caregivers. Further studies on larger cohorts are warranted that will perhaps also address treatment allocation according to grading.

Conflicts of Interest

The authors declare no conflicts of interest.