Saddle Nose in Granulomatosis With Polyangiitis (GPA) vs. Non‐GPA Patients With Septal Perforations

ABSTRACT

Objectives

Saddle nose deformity (SND) can be associated with nasal septal perforation (NSP). Both SND and NSP are observed in granulomatosis with polyangiitis (GPA) patients. This study aimed to compare the prevalence and severity of SND in GPA vs. non‐GPA patients with NSP.

Methodology

NSP patients who visited tertiary rhinology or facial plastic surgery clinics from January 2010 through December 2023 were grouped into GPA and non‐GPA cohorts. The area of NSP, SND incidence, and SND severity (Daniel and Brenner's classification) were compared between the cohorts.

Results

Of 168 patients identified with NSP, 18 had GPA and 150 had non‐GPA diagnoses. Twelve GPA patients and 10 non‐GPA patients had SND. The odds ratio for SND association in GPA patients with NSP vs. non‐GPA patients was 31.3 (95% CI: 9.5–102.9), and 19.4 (95% CI: 4.3–87.7) after controlling for the NSP area. GPA patients with SND had a larger NSP area (median 670 mm², IQR: 315.5–1061.3) compared to non‐GPA patients with SND (175.3 mm², 113.1–204.1; p < 0.001). The severity of SND was similar between GPA and the non‐GPA group (p = 0.09). The area of NSP did not predict the severity of SND (p = 0.17).

Conclusion

GPA diagnosis was associated with larger NSPs and a higher risk of SND compared to non‐GPA etiologies; however, the severity of saddling did not differ between GPA and non‐GPA etiologies. GPA is an independent risk factor associated with SND irrespective of the size of septal perforation. The NSP area did not predict the severity of SND.

Level of Evidence: 4.

This is a retrospective case‐control study comparing patients with nasal septal perforations with GPA vs. non‐GPA etiologies. GPA patients had larger perforations and 19 times higher odds of being associated with saddle nose deformity independent of perforation size. Severity of saddle nose was not different between the groups and perforation size did not predict the severity of saddling.

1. Introduction

Granulomatosis with polyangiitis (GPA) is a small to medium vessel vasculitis characterized by necrotizing granulomas. GPA usually involves a triad of systems: the upper airway, lungs, and the kidneys [1]. Upper airway manifestations are present in > 90% of patients and can include rhinorrhea, sinusitis, nasal septal perforation, and eventually deformities such as vestibular stenosis and saddle nose deformity (SND) [2]. Limited upper airway involvement is noted in 25% of GPA patients, which is characterized by the dominance of sinonasal symptoms as described [3]. GPA vasculitis classically affects the vessels of Kiesselbach's plexus of the anterior nasal septum to produce rhinorrhea and extensive nasal crusting symptoms commonly noted at presentation [4]. Disease progression with the destruction of the cartilaginous nasal septum and development of an anterior nasal septal perforation (NSP) is found in approximately 33% of GPA patients [5]. Perforations may enlarge and compromise dorsal and caudal support, precipitating retraction of the columella and a dorsal depression, or saddling. SND is found in approximately 30% of GPA patients (Figure 1) despite appropriate medical therapy [4].

FIGURE 1.

Photos and sinus computed tomography scan of a 38‐year‐old female with GPA. Frontal (a) and right lateral (b) views depicting a saddle deformity with over rotation and deprojection of the nasal tip (type III deformity). Sagittal scan view (c) demonstrates the 16 mm (length) × 13 mm (height) perforation involving the quadrangular cartilage posteriorly (red dotted circle). The distance from the perforation to the dorsal profile line is 21 mm. Crusting is present around the perforation visible in the perforation lumen.

SND results from the loss of cartilaginous (quadrangular septal cartilage and upper lateral cartilages) and/or bony support (bony septum and nasal bones) of the nasal dorsum [6]. As a result, facial esthetics, as well as nasal function, are compromised due to the collapse of the dorsum and the internal nasal valve. The general overall prevalence of SND is not well established due to its rarity and association with specific conditions that include nasal trauma, overzealous rhinoplasty or septoplasty maneuvers, and medical conditions including systemic inflammatory diseases (GPA, relapsing polychondritis, sarcoidosis, systemic lupus erythematosus) and infections (tuberculosis, syphilis) [7].

The aim of this study was to determine the risk of SND in NSP patients with and without GPA and to determine the relative odds of SND between these groups. Additionally, we aim to compare the size of NSPs in these groups and investigate the relationship between NSP size and the degree of SND.

2. Methodology

After approval of this study by the Institutional Review Board (IRB—24‐000840), adult patients who presented to the rhinology or facial plastics surgery clinic of a tertiary care center (from January 2010 until December 2023) for management of NSP were identified using the Informatics for Integrating Biology and the Bedside (i2b2) software provided by Mayo Clinic. The etiology of NSP was identified by a review of individual medical records and used to divide patients into two groups: with and without GPA. Patients with a history or current diagnosis of sinonasal tumors were excluded from this study.

Information on patients' age, sex, perforation dimensions, and presence or absence of a SND were noted by reviewing the clinical notes, sinus computed tomography (CT) scans, and patient photographs. Perforation dimensions were noted from the operative note, or the in‐clinic endoscopic examination note when available or otherwise measured from the patients' sinus CT scans [8]. Clinical pictures were used to assess the severity of SND according to Daniel and Brenner's grading (Type I to V, higher is worse) [6]. As NSPs are almost always elliptically shaped, the formula for calculating the area of an ellipse was used for calculating the area of NSPs (π × r₁ × r₂, where r₁ = length of NSP/2 and r₂ = height of NSP/2) [8]. Patients' age, sex, area of NSP, incidence of SND, and type/grade of SND were compared between the GPA and non‐GPA groups.

STATA/BE 18.0 software was used for statistical analysis. The odds ratio was used to compare the incidence of SND in GPA vs. non‐GPA patients. For calculating the odds of the presence of SND between the two groups independent of the area of NSP, the latter was controlled using logistical regression. Fischer's exact test was used to compare the sex, and the Wilcoxon Rank Sum test was used to compare the age, area of NSP, and type of SND between the patients of the two groups. Ordinal logistical regression was used to assess if the NSP area predicted the SND grade in patients of both groups having SND. A p‐value of < 0.05 was considered significant.

3. Results

Eighteen patients with GPA and 150 without GPA were identified with NSPs. Twelve patients in the GPA (66.7%) and 10 in the non‐GPA cohort (6.7%) had SNDs. Age (p = 0.95) and sex (p = 0.08) distribution were not significantly different between the cohorts. In patients with an SND in the non‐GPA group, etiologies for NSPs included prior septal surgery in 5, chronic nasal spray use in 1, cocaine abuse in 2, and button battery foreign body in 1. The median area of NSP in the GPA group was 531 mm² [Interquartile range (IQR): 239–1017], which was significantly larger than the non‐GPA group, where it was 126 mm² (35–177); p < 0.001. When calculating only for the patients with an SND, the median area of NSP in the GPA group was 670 mm² (315.5–1061.3), which was again larger than the non‐GPA group [175.3 mm² (113.1–204.1)]; p < 0.001. The odds ratio for the presence of SND in the GPA group vs. the non‐GPA group was 31.3 [95% Confidence interval (CI): 9.5–102.9]. After controlling for the area of NSP using logistical regression, GPA patients had 19.4 times higher odds (95% CI: 4.3–87.7) of developing an SND than patients in the non‐GPA group; p < 0.001. There was no significant difference in the severity of SND between the GPA and non‐GPA cohorts (p = 0.09). The ROC (Receiver Operating Characteristic) curve analysis was conducted to evaluate the association between the area of NSP and the presence of SND in patients with GPA and non‐GPA etiologies. In the GPA cohort, the area under the curve (AUC) was 0.63 (95% CI: 0.31–0.95), and in the non‐GPA cohort, AUC was 0.69 (95% CI: 0.54–0.84). Although the ROC curve analysis suggested a better ability to distinguish between the area of NSP and the presence of SND in the non‐GPA cohort, overall, both cohorts had poor discriminative power. Using ordinal logistical regression, we noted that the area of NSP did not predict the severity of SND; p = 0.17. Clinical details of patients are provided in Table 1.

TABLE 1.

Clinical details of patients with nasal septal perforation included in the study.

	GPA cohort	Non‐GPA cohort	p
Total no. of patients	18	150
Mean age ± SD (yrs.)	52.4 ± 15.6	52.3 ± 16.9	0.95
Sex		[not disclosed in 1]	0.08
M	4 (22.2%)	68 (45.3%)
F	14 (77.8%)	81 (54%)
Median NSP area (mm2), IQR	531, 239–1017	126, 35–177	< 0.001
No. of patients with SND	12	10
Mean age ± SD (yrs.)	53.4 ± 16.4	38 ± 14.5	0.052
Sex			0.07
M	2 (16.7%)	6 (60%)
F	10 (83.3%)	4 (40%)
Median NSP area (mm2) in patients with SND, IQR	670, 315.5–1061.3	175.3, 113.1–204.1	< 0.001
Type of SND (Daniel and Brenner's classification [1])	[Unknown in 1]		0.09
I	3 (25%)	5 (50%)
II	1 (8.3%)	2 (20%)
III	7 (58.3%)	2 (20%)
IV	0	0
V	0	0

Abbreviations: GPA, granulomatosis with polyangiitis; IQR, interquartile range; M/F, male/female; NSP, nasal septal perforation; SD, standard deviation; SND, saddle nose deformity.

4. Discussion

GPA patients with nasal septal perforation were found to have ~31 times higher odds of developing a SND compared to the patients with non‐GPA etiologies. NSPs were noted to be larger in the GPA cohort. Still, when controlling for the area of NSP, GPA patients with NSPs had ~19 times higher odds of being associated with an SND compared to patients with non‐GPA perforation etiologies, implying that GPA diagnosis is an independent risk factor associated with SND. Interestingly, the area of NSP did not predict the presence or severity of SND. The severity of SND in GPA patients with larger NSPs did not significantly differ from the non‐GPA cohort.

The necrotizing granulomatous inflammation associated with GPA involves Kiesselbach's plexus of the cartilaginous nasal septum [4]. Inflammation, injury, and necrosis can lead to a septal perforation that may progress in size with ongoing vasculitis and cartilage injury. Perforation enlargement and the loss of dorsal septum structural integrity weaken the dorsum's support to produce SND [7]. Progression to subtotal perforation and severe saddling can be seen in GPA patients (Figure 1). This was also observed in our study where GPA patients had larger perforations and markedly higher odds of being associated with an SND.

After controlling for the area of NSP, GPA patients continued to have approximately 19 times higher odds of being associated with an SND. The area of NSP was also found to be poorly associated with the presence of SND, as shown by the ROC curve analysis. Additionally, the area of NSP did not predict the severity of SND. These observations suggest that although NSP size can be a key factor in the onset and severity of SND, factors beyond structural septal loss may contribute to the frequent saddling observed in GPA patients. Saddling can be observed in GPA patients with an intact nasal septum or a very small perforation (Figure 2). Vasculitis and granulomatous inflammation can reduce tissue perfusion [1] causing dissolution of the septal cartilage that compromises the dorsal support independent of the presence of a septal perforation. However, this hypothesis lacks validation in existing literature. Akiyama et al. [9] observed that GPA patients with midline cartilaginous and bone lesions in the head and neck region were highly efficient in generating neutrophil extracellular traps (NETs), where trapped neutrophils released alarmins, further activating tissue‐destructive monocytes. This could also point toward more extensive septal cartilage destruction than what is observed in the form of a NSP. We have noted a loss of cartilage integrity extending from the perforation's superior margin to the depressed septal‐upper lateral cartilage junction in some GPA patients when attempting surgical repair. Presumably, this inflammatory process can progress to the keystone region (K‐area). The K‐area represents the vertical region posterior to the caudal septum and inferior to the bony‐cartilaginous junction that is critical for dorsal support [10]. Loss of structural integrity in this area due to a large septal defect or the inflammatory dissolution of cartilage may increase the risk for substantial saddling compared to other areas of the septum, as illustrated in Figure 1. Therefore, NSP size and location, as well as the strength of the septal quadrangular cartilage, may be the factors that determine the precipitation and severity of an SND. The dissolution of septal cartilage beyond the margins of septal perforation can also give a false impression of a smaller area of septal cartilage loss. Hence, palpation with a cotton tip applicator may provide clinically relevant information on perforation size and septal structural integrity beyond the perforation margin. This information can be vital in improving our understanding of the impact of the loss of septal cartilaginous support in the development of an SND.

FIGURE 2.

Parasagittal view of a sinus CT scan of two different GPA patients showing (a) small sub‐centimetric perforation with 4 mm length and 3 mm height (red dotted circle) situated in the posterior part of quadrangular septal cartilage with a Type I saddle nose deformity (white arrow), and (b) total nasal septal perforation without a saddle nose deformity.

Saddle deformity may not occur in patients with large NSPs, including those of inflammatory GPA or cocaine etiology [6] (Figure 2b). In addition to the cartilaginous septum, the bony septum, upper and lower lateral cartilages, and nasal bones also contribute to nasal dorsal support. These other structures are typically spared of GPA‐associated inflammation [11]. Their contribution to supporting the dorsum and preventing saddling in the event of loss of cartilaginous septal support is not known. This information might be key to understanding why some GPA patients do not develop an SND even after losing the complete cartilaginous septum, whereas a few with an apparently intact septum do.

This is the first study to compare the characteristics of GPA‐associated SND and nasal septal perforations with non‐GPA etiologies. Although previous studies have reported an association of GPA with NSP and SND [11], a comparison with non‐GPA pathologies has not been done. This study helps to improve our understanding of GPA‐associated NSP and SND. Future studies may provide further information to understand the impact of GPA‐associated septal inflammation on septal cartilage structural integrity and the other structures contributing to dorsal support and SND.

Our study has several limitations, including those inherent to a retrospective chart review. Due to the overall low incidence of SND, all patients with non‐GPA etiologies have been grouped in a single cohort against the cohort of GPA patients to allow for a meaningful comparison. This grouping prohibits us from comparing the characteristics of interest in our study between GPA patients and patients with any particular non‐GPA etiology. Patients with perforations and external deformities due to cocaine inhalation are underrepresented in the data. We do not routinely obtain photos at the time of the initial presentation, and many cocaine patients do not return for subsequent evaluation following treatment recommendations. GPA is a rare condition; therefore, the sample size is small in the GPA cohort relative to the size of the non‐GPA cohort, creating heterogeneity between the groups for a direct comparison. With a low sample size of the GPA cohort and the rarity of SND in patients with non‐GPA etiologies, even fewer patients had SND in both cohorts. This might render most of our observations and analyses underpowered; however, this is the data of the last 14 years from a tertiary care center with a large nasal septal perforation practice.

5. Conclusions

GPA diagnosis was associated with larger nasal septal perforations and a higher risk of SND compared to non‐GPA etiologies. However, the severity of saddling did not differ between GPA and non‐GPA etiologies. GPA is an independent risk factor associated with SND irrespective of the NSP size. We found that the area of NSP does not appear to predict the severity of SND. Further studies are needed to investigate the impact of the presence and size of NSP on the formation and severity of SND. Assessment of the impact of GPA vasculitic inflammation on the strength of the cartilaginous septum may improve our assessment and treatment of patients with saddling and small or absent septal perforation.

Disclosure

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Kumar N., Gomes P. L., Jett S., et al., “Saddle Nose in Granulomatosis With Polyangiitis (GPA) vs. Non‐GPA Patients With Septal Perforations,” Laryngoscope Investigative Otolaryngology 10, no. 4 (2025): e70217, 10.1002/lio2.70217.