Epidemiology of pheochromocytoma and paraganglioma: population-based cohort study

Introduction

Pheochromocytomas and paragangliomas are rare tumors derived from chromaffin tissue with the potential for catecholamine secretion. Though rare, these are important to recognize because, if left untreated, tumors may lead to high blood pressure, cardiovascular complications, and even death (1, 2). The prognosis, however, is typically excellent in cases that are medically stabilized and resected with surgery (3). Experienced multidisciplinary teams are ideally suited to manage patients perioperatively and to provide postoperative follow-up to monitor for recurrence, distant metastasis, or associated tumors (4, 5).

Disease surveillance is a critical first step for healthcare professionals and policy makers to allocate appropriate resources and to improve access for specialized treatment for affected patients. However, despite the significant morbidity and mortality associated with pheochromocytoma and paraganglioma, there are no mechanisms in place to routinely monitor for this disease in the population. An accurate and updated understanding of the epidemiology of pheochromocytoma and paraganglioma is vital, not only to inform risk management, but to provide the needed data to evaluate current practice and policy.

Little is known about the epidemiology of this condition. Previous studies have been limited in their generalizability as these have commonly been conducted in selected populations, mostly representative of people of white ethnicity (6, 7, 8, 9, 10), those living in catchment areas near highly specialized tertiary and quaternary care centers (10, 11), and many based on data collected over half a century ago (7, 8, 9, 11). Addressing this, we conducted a cohort study from an ethnically and sociodemographically diverse population in western Canada within the last decade, identifying incident cases of pheochromocytoma and paraganglioma using high-quality clinical records. Furthermore, we developed and validated several algorithms for case detection based on laboratory and administrative data to facilitate long-term disease surveillance in the population.

Methods

Results

During the 7 years of observation, there were 5 196 368 residents of Alberta. Of these, 15 113 people received biochemical work-up to screen for pheochromocytoma or paraganglioma. A total of 2158 individuals had at least one physician encounter for a potential diagnosis of pheochromocytoma or paraganglioma based on codes derived from provider claims, emergency department visits, and hospital discharge. After review of the corresponding charts, pathology, and autopsy records, 239 people were verified to have pheochromocytoma and/or paraganglioma. Nearly all diagnoses were confirmed using surgical pathology, tissue biopsy, or autopsy, collectively accounting for 237 (99.2%) of the cases (Table 1).

Of the 239 individuals with pheochromocytoma and/or paraganglioma, the median age at diagnosis was 55 years (Table 1). More than half of affected individuals were female and 5 (2.1%) were pregnant at the time of diagnosis. While most people had a single tumor, 9 (3.8%) individuals had multiple primaries. In total, 251 tumors were present with nearly half of these (49.4%) being pheochromocytomas. Most of the remaining lesions were situated in the head and neck area (37.1%). The median size of the detected lesions was 3.2 cm in maximum dimension. Males tended to have larger tumors compared to females (median 3.7 cm vs 3.0 cm, respectively; P = 0.004).

Incidence of pheochromocytoma and paraganglioma

The overall incidence was 0.66 cases per 100 000 people per year (Table 2). Approximately half of the individuals had a defining diagnosis of pheochromocytoma with an incidence of 0.32 cases per 100 000 people per year; the other half had paraganglioma with an incidence of 0.33 cases per 100 000 people per year. These rates remained broadly similar throughout the entire observation period (Supplementary Table 3). The incidence of pheochromocytoma and paraganglioma was generally higher in females compared to males in nearly every age group except in children (Fig. 1). When examined according to age bands, children (aged 19 years and younger) had the lowest rates of pheochromocytoma (0.37 and 0.19 cases per 100 000 people per year for males and females, respectively) and paraganglioma (0.18 and 0.19 cases per 100 000 people per year for males and females, respectively). Comparatively, adults (aged 20 years and older) had a higher incidence of pheochromocytoma (3.27 and 3.84 cases per 100 000 people per year for males and females, respectively) and paraganglioma (2.53 and 4.83 cases per 100 000 people per year for males and females, respectively). With increasing age, there was a linear rise in the incidence of pheochromocytoma and paraganglioma up to 60–79 years (8.85 and 14.68 cases per 100 000 people per year for males and females, respectively), which was then followed by a decline in the frequency of disease detected in those aged 80 years and older.

Incidence of pheochromocytoma or paraganglioma (Panel A), pheochromocytoma (Panel B), and paraganglioma (Panel C) according to age and sex strata. The population census from July 1, 2018 was used to represent the average size and demographics of the province over the 7-year period. Men are represented in blue and women in red.

Citation: European Journal of Endocrinology 184, 1; 10.1530/EJE-20-0628

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Incidence of pheochromocytoma or paraganglioma (Panel A), pheochromocytoma (Panel B), and paraganglioma (Panel C) according to age and sex strata. The population census from July 1, 2018 was used to represent the average size and demographics of the province over the 7-year period. Men are represented in blue and women in red.

Citation: European Journal of Endocrinology 184, 1; 10.1530/EJE-20-0628

• Download Figure
• Download figure as PowerPoint slide

Incidence of pheochromocytoma or paraganglioma (Panel A), pheochromocytoma (Panel B), and paraganglioma (Panel C) according to age and sex strata. The population census from July 1, 2018 was used to represent the average size and demographics of the province over the 7-year period. Men are represented in blue and women in red.

Citation: European Journal of Endocrinology 184, 1; 10.1530/EJE-20-0628

• Download Figure
• Download figure as PowerPoint slide

Table 2

Incidence of people diagnosed with pheochromocytoma (PHEO) and/or paraganglioma (PG) between April 1, 2012 and March 31, 2019.

	No. of people diagnosed*	Population at risk	Incidence per 100 000 people†
Tumour type
PHEO or PG	239	5 196 368	0.66
PHEO	118		0.32
PG	121		0.33

*The first occurrence of a PHEO or PG in an individual was used to calculate incidence. The occurrence of metachronous primary tumors, tumor recurrences, and metastases were not considered in the incidence calculation.; ^†Calculated incidence from 2012 to 2019 may be an underestimation of the true rate because individuals who left the province were assumed to have remained disease-free.

Discussion

In this study, the annual incidence of pheochromocytoma and paraganglioma was observed to be 0.6–0.7 cases per 100 000 people (or 6–7 cases per million people). Of note, the overall incidence calculated from 2012 to 2019 was likely an underestimate of the true rate (i.e. 0.7–0.9 cases per 100 000 people, as observed from year to year) because individuals who left the province were assumed to have remained disease-free for all 7 years. While admittedly rare, these tumors still occur more frequently than commonly believed, challenging previous dogma suggesting this condition is only present in around one in a million, and that no more than one case of pheochromocytoma or paraganglioma will be admitted to a large general hospital per year (10, 17). Here, at least a dozen cases were treated per year in each of the two largest tertiary care hospitals in the province (both serving a catchment of over 2 million people). We also found that laboratory and administrative data were relatively accurate in estimating incidence in our population, even with the high variability in the anatomical location, hormonal activity, and clinical presentation of these tumors, and therefore these serve as potential alternatives to facilitate ongoing surveillance.

Apart from the present study and another from the Netherlands (6), knowledge of the epidemiology of pheochromocytoma and paraganglioma is largely limited to reports from the 1960s to 1980s, as contemporaneous studies performed within the last 25 years are otherwise lacking. Historical incidence rates have been considerably lower than those observed here with most studies suggesting an annual incidence of approximately 0.2 cases of pheochromocytoma and paraganglioma per 100 000 people per year (7, 8, 10). These lower rates are likely subject to important differences in how cases were ascertained in older studies (e.g. with some relying on clinician recall (8, 9), certain studies using disease registries that may have been incomplete (8, 10), and others only sampling hospital-based or surgical records) (6, 7, 9). In contrast, using a combination of multiple high-quality data sources with province-wide linkage, we were able to ensure complete capture of all clinical encounters in a geographically well-defined population, thus minimizing the chance of missed cases. While we observed a slightly lower rate of pheochromocytoma and paraganglioma in our diverse population compared to those reported from the Mayo Clinic in Rochester, Minnesota (i.e. 0.95 cases per 100 000 people per year) (11) and the Netherlands (i.e. 0.94 cases per 100 000 person per year) (6), this may simply reflect the fact that these later studies were conducted in selected populations at greater risk of disease (i.e. the Mayo Clinic being a high-referral center and the people of the Netherlands potentially harboring a greater rate of associated genetic abnormalities) (18). Moreover, we found that pheochromocytomas accounted for only half of all tumors and this was proportionately less than expected. Most likely, this was because head and neck paragangliomas (e.g. jugulotympanic tumors, carotid body tumors, and glomus vagale tumors), which have not been systematically included in previous epidemiological studies but were included here, accounted for more than one-third of all our cases. Our findings also suggest that most contemporary cases of pheochromocytoma and paraganglioma are detected during life rather than after death (as was the case in the past where up to half of diagnoses were made post-mortem) suggesting an increased global awareness of these tumors and improvements in diagnostic practices over time (6, 8, 19, 20, 21). Finally, we found that among children, there was a slight male preponderance of pheochromocytoma and paraganglioma (22). In adults, incidence generally climbed with age with females more commonly affected than males (6, 8, 10). Altogether, the present study is consistent with previous work, and provides a more refined estimate of the true incidence of pheochromocytoma and paraganglioma in an unselected population.

Given the high cost and labor needed to perform manual chart review to confirm each case of pheochromocytoma and paraganglioma in the population with the traditional method of determining incidence, we recognized the need for a more sustainable approach. Accordingly, in a universal healthcare system, routinely collected laboratory and administrative data can be leveraged for population-based evaluations of disease burden (13, 15). We demonstrated that there was reasonable diagnostic agreement between the case definitions developed from laboratory and administrative data, and our reference standard. The principal advantage of using laboratory and administrative data was that these were considerably less costly to acquire and more efficient to analyze compared to manual chart review. The main disadvantage of these data related to the rarity of pheochromocytoma and paraganglioma, leading to a high false-positive rate (23, 24, 25), as was directly reflected by the modest sensitivities and PPVs of the derived algorithms, and correspondingly wide 95% CIs. Still, while the aforementioned data sources were somewhat less accurate, the predicted estimates of pheochromocytoma and paraganglioma incidence appeared broadly similar to those expected, and therefore generally sufficient for the purposes of disease surveillance (26, 27, 28). This method of surveillance not only helps to quantify disease burden, but also defines a population where relevant process and outcome measures can be explored (e.g. examination of regional variations in treatment). Though not directly used in clinical care, such algorithms have been successfully applied to identify other medical conditions to facilitate population-based research and largescale surveillance (13, 15, 29).

There are many strengths of our study (i.e. we identified a large cohort drawn from an ethnically and sociodemographically diverse population with complete capture of all clinical encounters; by using multiple linked high-quality data sources, we were less susceptible to responder biases; and, we were able to validate multiple algorithms based on laboratory and administrative data for case detection with an independent reference standard). Our study was subject to certain limitations. First, it is possible that some people with pheochromocytoma or paraganglioma could have been overlooked if an individual had no relevant healthcare encounters. Addressing this, we reviewed all available autopsy reports, but it was impossible to ensure that there was complete case ascertainment, as the provincial autopsy rate was only around 7% (https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1310071601&pickMembers%5B0%5D=1.10; accessed June 5, 2020). While estimated incidence rates may have been subject to some degree of misclassification, this would have likely been small. While it was not possible to absolutely confirm the true negative rate in the population, it was unlikely that clinically significant cases were missed. Second, we collectively grouped pheochromocytoma and paraganglioma together, even though important clinical differences may exist with respect to hormonal function, disease trajectory, and prognosis. As such, we tried to classify tumors according to their anatomical locations. We were unable to reliably determine, particularly in the case of paraganglioma of the head and neck region, if these arose from the sympathetic or parasympathetic chains. Nevertheless, non-functional tumors are still clinically important to detect as they pose a risk of local invasion, mass effect, and even distant metastasis. Finally, we did not have some information that would have been relevant for our study, including individual ethnicity (though population census profiles are available; https://www12.statcan.gc.ca/census-recensement/2016/dp-pd/prof/details/page.cfm?Lang=E&Geo1=PR&Code1=48&Geo2=PR&Code2=01&Data=Count&SearchText=alberta&SearchType=Begins&SearchPR=01&B1=All&TABID=1; accessed June 5, 2020), related comorbidities, and the presence of germline mutations. These data are relevant because at least 18 susceptible genes have been reported to increase the risk of pheochromocytoma and paraganglioma (e.g. NF1, RET, VHL, SDHA, SDHB, SDHC, SDHD, etc.) with some of these potentially occurring more frequently in certain ethnic groups than others due to the presence of founder mutations (30). Further study is needed to improve our understanding of the association between ethnicity, pheochromocytoma, and paraganglioma so that the risks of developing disease can be better defined and personalized surveillance strategies can be established for affected kindreds.

In conclusion, the frequency of pheochromocytoma and paraganglioma in the general population of western Canada was unexpectedly high relative to many other jurisdictions (7, 8, 10), posing important clinical implications. As pheochromocytoma and paragangliomas are commonly associated with high rates of inherited predisposition, our results may indicate an even higher burden of undiagnosed tumors in asymptomatic and unscreened family members. In light of the significant risks to morbidity and mortality posed by pheochromocytoma and paraganglioma, accurate monitoring is needed for risk management, to identify important prognostic factors, to guide policy for relevant care pathways, and to meaningfully inform the use of clinical resources for affected patients. Accordingly, population-based data represent valuable assets for those seeking to evaluate the burden of disease, delivery of care, and outcomes for affected patients.

Supplementary materials

This is linked to the online version of the paper at https://doi.org/10.1530/EJE-20-0628.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this study.

Funding

Alexander A Leung is supported by the Hypertension Canada New Investigator Award.