Re‐evaluating the prevalence and factors characteristic of catecholamine secreting head and neck paragangliomas

Abstract Introduction We sought to characterize the prevalence and factors characteristic of head and neck paragangliomas (HNPGLs) that secrete catecholamines to inform best practices for diagnosis and management. Methods This was a retrospective cohort study from 2000 to 2020 at a single‐institution tertiary centre. One‐hundred fifty‐two patients (182 tumours) with HNPGLs with at least one measurement of urine or plasma catecholamines and/or catecholamine metabolite levels prior to treatment were included. We differentiated and characterized those patients with increased level(s) of any nature and those with ‘clinically significant’ versus ‘clinically insignificant’ catecholamine production. Results Thirty‐one (20.4%) patients had increased catecholamine and/or catecholamine metabolite levels. In most patients, these levels were ≤5‐fold above the upper limit of the reference range. Four of these 31 patients with increased levels were ultimately found to have an additional catecholamine secreting mediastinal paraganglioma or pheochromocytoma. Fourteen of 31 patients with HNPGL were deemed clinically significant secretors of catecholamines based on hyper‐adrenergic symptoms and/or profound levels of normetanephrines. This cohort was enriched for patients with paragangliomas of the carotid body or cervical sympathetic chain and those with SDHB genetic mutations. Ultimately, the prevalence of clinically significant catecholamine secreting Hangs was determined to be 9.2% and 7.7% based on a per‐patient and per‐tumour basis, respectively. Conclusions The rate of catecholamine excess in the current cohort of patients with HNPGLs was higher than previously reported. Neuroendocrine tumours of any anatomic subsite may secrete catecholamines, although not all increased laboratory level(s) are indicative of clinically significant catecholamine secretion causing symptoms or warranting adrenergic blockade.


| INTRODUC TI ON
Head and neck paragangliomas (HNPGLs) are rare tumours derived from paraganglial cells within autonomic ganglia of the carotid body (CBP), vagus nerve (VP), jugular bulb (JP), Jacobsen's nerve of the middle ear (TP) or cervical sympathetic chain (SCP). 1 Due to their cell of origin, HNPGLs have the potential to actively synthesize and secrete catecholamines with potentially deleterious systemic effects.
Patients with catecholamine secreting HNPGLs may present with symptoms of catecholamine excess, including sustained or intermittent hypertension and tachycardia, cardiac palpitations, diaphoresis and/or pallor. Regardless of symptomatology, failure to identify and treat catecholamine excess may cause significant morbidity and even mortality in these patients, particularly those undergoing surgery. 2 As a result, contemporary clinical practice guidelines recommend biochemical testing of urine or plasma catecholamines and metabolites, usually metanephrine and normetanephrine levels, for all patients with newly diagnosed HNPGLs. 3 The prevalence of catecholamine secreting HNPGLs is typically low (approximately 3%-4% of tumours). 4,5 However, this estimate is based primarily on data from limited series published 20 years ago. More recent evidence suggests that the rate of catecholamine secreting HNPGLs may exceed 10%. 6 Recently, biochemical and genetic characteristics of HNPGLs, including standard treatment paradigms have been vastly transformed. 7,8 There now exists a clear gap in the literature regarding the true rate and factors characteristic of catecholamine secreting HNPGLs, particularly those whose functional status may pose significant challenges for peri-operative management.
Here, we report a large series of patients with HNPGLs with the primary aim of characterizing the prevalence and features of HNPGLs that secrete catecholamines at a level significant enough to cause symptoms or warrant consideration of adrenergic blockade, herein termed 'clinically significant'. In an era of increasingly personalized treatment for these tumours, our data may inform contemporary, best practices for biochemical screening and multi-disciplinary management of HNPGLs.

| ME THODS
This was a retrospective analysis of a prospectively maintained clinical database of patients with HNPGLs presenting to our institution for evaluation and management between 2000 and 2020. 9 Inclusion criteria for this study were as follows: (1) radiographically confirmed isolated or multi-focal HNPGL; (2) previously untreated HNPGL tumour(s); and (3) at least one laboratory measurement of urine or plasma catecholamine or catecholamine metabolite levels prior to treatment onset.
Urine measurements included analysis of 24-h excretion of fractionated normetanephrines and metanephrines, vaniyllmandelic acid (VMA), norepinephrine, epinephrine and dopamine via standard clinical assays. 10 Similarly, plasma measurements included analysis of fractionated normetanephrines and metanephrines, VMA, norepinephrine, epinephrine and dopamine, as described. 10 As expected, laboratory reference ranges were not uniform due to the twenty- year study period, differences in clinical assays, and few patients with laboratories from other institutions that were not repeated upon presentation. As such, we recorded absolute laboratory levels and calculated a normalized 'per cent of reference range' level for each measurement as follows: [(absolute level-lower bound of reference range)/upper bound of reference range] x 100. Laboratory assessments were considered increased when the per cent of reference range level exceeded 100%. 11 As previous authors have posited, it is a clear oversimplification to characterize HNPGLs as simply functional or not. Rather, HNPGLs exhibit a continuum of hormonal activity influencing clinical presentation and need for hemodynamic management. 12,13 Thus, the primary goal of this study was to investigate the rate and Operational definitions for other recorded clinical variables are as follows: hyper-adrenergic symptoms at first presentation were defined as explicit documentation of sustained or intermittent palpitations, tachycardia, diaphoresis, and/or tremors or new-onset hypertension in conjunction with at least one of these other symptoms. Hypertension at first presentation was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg.
Tachycardia at first presentation was defined as resting heart rate >100 beats per minute.
Statistical comparisons between groups were made with chisquare test and Student's t test for categorical and continuous variables, respectively. All statistical tests were two-tailed and performed with SPSS Version 27 with a p ≤ .05 as the threshold for statistical significance. This study was deemed exempt from informed consent by the University of Michigan Institutional Review Board (IRB).

| RE SULTS
Our study cohort consisted of 280 patients with HNPGLs comprising 318 discrete tumours. There were significant differences evident among patients in whom laboratories were drawn/ documented (n = 152, 182 tumours) versus not (n = 128, 136 tumours). Specifically, patients in the 'labs drawn' group were younger overall and more likely to endorse hyper-adrenergic symptoms at presentation. HNPGL tumour subsite(s) also differed, with more JP and multi-focal HNPGLs and fewer TP in the 'labs drawn' group overall (Table 1).
Within the 'labs drawn' group, the specific laboratory assessments of catecholamine and/or catecholamine metabolite levels varied considerably ( Figure 1A). Twenty-four-hour urinary dopamine excretion was the least commonly employed test (n = 12, 7.9%) while plasma normetanephrines and metanephrines (n = 90, 59.2%) were most frequently assessed in our cohort. Over the course of the study period, we saw a modest increase in percentage of patients with HNPGL who had lab(s) drawn at first presentation ( Figure 1B).
Further, we saw a significant but opposite trend in the use of urine and plasma normetanephrine and metanephrine assessments over time (p < .01 for both trends, Figure 1C).
The median and interquartile range of all laboratory measurements are provided in Table S1. In total, 31 (20.4%) of 152 patients had one or more laboratory assessments showing increased catecholamine or catecholamine metabolite levels. In most patients with HNPGL, these levels were ≤5-fold above the upper limit of the reference range, though a few individuals had profound laboratory levels of ≥10-fold ( Figure 2, Table S2).   In total, 14 patients in our cohort were determined to have evidence of clinically significant catecholamine secretion from their tumours ( Table 2). This small cohort was particularly enriched for patients with tumours of the carotid body (CBP) and cervical sympathetic chain (SCP) as well as those with pathogenic SDHB mutations.
Ten of these 14 patients were started on α-and/or β-blockade after laboratory assessments were completed. We could not determine whether blockade was initiated in the remaining four patients due to insufficient clinical documentation or limited follow-up duration. In summary, of 152 HNPGL patients with 182 total tumours, clinically significant catecholamine secretion was shown in 9.2% and 7.7% on a per-patient and per-tumour basis, respectively.
Lastly, we sought to evaluate the sensitivity and specificity of hyper-adrenergic symptoms at first presentation for both increased laboratory level(s) and clinically significant catecholamine secretion.
As expected, the sensitivity for both outcomes was quite low (44.8% and 44.4%, respectively). Conversely, specificity was much higher at 83.3% and 79.8%, respectively (Figure 4).  biochemical screening for newly diagnosed HNPGLs was also more common in those patients who endorsed such symptoms potentially attributable to hormonally active tumours.
As a rapid and easy screening tool, we hypothesized that more uniform assessment of hyper-adrenergic symptoms at first presentation may be warranted to increase detection rates of catecholamine secreting HNPGLs. We found that sensitivity of such symptoms for any increased laboratory level(s) and clinically signif- Based on our vast institutional experience, it is evident that increased laboratory level(s) indicative of catecholamine excess may not always lead to significant changes in clinical management. A recent paradigm shift, towards watchful waiting and/or radiation to avoid surgical morbidity for benign tumours was evident even in our patients with clinically significant catecholamine secreting HNPGLs. As such, biochemical screening for catecholamine excess has become essential for blood pressure and heart rate control with adrenergic blockade during the period of watchful waiting or radiation. For those patients treated surgically, our data reiterate the strong indication for biochemical screening in pre-surgical workup to avoid rare but catastrophic peri-operative complications. 23 A measurable percentage (8.6%) of our cohort had clinically insignificant catecholamine secretion as we defined it. These patients had increased laboratory levels that were irreproducible, non-specific or minor and not associated with hyper-adrenergic symptoms. When patients first present to an otolaryngologisthead and neck surgeon, consultation with an endocrinologist with expertise in such tumours is ideal to help interpret laboratory levels and advise on appropriate next steps in diagnosis and management.
While initiation of adrenergic blockade may not be immediately required for these individuals, the questionable catecholamine excess may prompt additional testing, imaging or genetic screening.

| CON CLUS IONS
The rate of catecholamine excess in patients with HNPGLs may be higher than previously thought. Tumours of any anatomic subsite may secrete catecholamines, although not all increased laboratory level(s) are indicative of clinically significant catecholamine secretion causing symptoms or warranting adrenergic blockade. Our series provides a comprehensive, contemporary update on biochemical profiles of HNPGL in an era of evolving diagnostic and management standards for these tumours.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to disclose relevant to this manuscript.

AUTH O R CO NTR I B UTI O N
We certify that all authors have met the following criteria for author-