Adrenal insufficiency (AI) results in loss of key steroid hormones that regulate stress responses, glucose metabolism, maintenance of blood volume, vasomotor tone, muscle and nerve function, and numerous other homeostatic effects. Causes of AI include primary adrenal insufficiency (PAI) due to autoimmune attack on the adrenals, infection, or injury; or secondary/central adrenal insufficiency (SAI), due to interruption of the hypothalamic-pituitary axis (HPA) from local tumor, inflammatory damage or suppression by exogenous glucocorticoids. If not identified and treated in a timely fashion, AI can result in life threatening consequences , , , . Little is known about diagnostic accuracy or the prevalence of AI in the US and there are no published US cohorts.
PAI is uncommon, with European prevalence approximately 200 per million or 0.02% , , . Using these estimates there could be 55,000–100,000 affected individuals in the US, but there are no sources to confirm this estimate and the US ethnic and racial distribution is distinctly different from Europe. Autoimmune disease has supplanted infections as the most common cause of PAI in European studies, but this may not be true within the US as both exposures and genetic risks may differ.
Major issues persist in the care of AI patients based on published reports of delays in diagnosis , , persistent excess mortality , , , , , impaired quality of life , , and episodes of adrenal crisis . Disease management strategies for AI in the US may differ from Europe and may vary by region within the US. Issues in disease management typically include: matching dosage of cortisol replacement to individual needs, avoiding excess cortisol effects, patient education for self-management and prevention of adrenal crises.
In order to address these needs, appropriate patient populations are needed for study, but the US healthcare system does not facilitate the creation of rare disease research cohorts. This is a problem for a disease like PAI that is rare, diagnostically difficult and is associated with occult death. Social media and patient advocacy groups provide alternative mechanisms to recruit subjects for research in rare diseases. We identified a patient registry of adrenal insufficiency patients through the National Adrenal Diseases Foundation (NADF) that offered information about the current status of diagnosis, treatment and patient safety in the US. We limited the analysis to patients with self-reported PAI.
Materials and methods
The NADF developed an on-line patient registry in collaboration with the American Autoimmune Related Diseases Association, Inc. (AARDA) in July of 2015. Participants in the registry were informed prior to completing the survey that their aggregated, de-identified data would be made available to researchers for analysis. This analysis of the NADF patient registry data was reviewed and approved by the National Jewish Health Institutional Review Board.
To focus on the state of PAI in the US, we excluded non-US residents from the cohort. Of 941 records in the registry, 759 subjects were from the US. An additional 46 subjects who were under age 20 or failed to report their age were excluded. Of the remaining 713, 541 participants met the case definition for PAI (reported having a physician diagnosis of “autoimmune or primary” adrenal insufficiency or Addison’s disease and reported taking glucocorticoid replacement in appropriate doses for these conditions).
Data and variables
Demographics included age, gender, race/ethnicity, marital status, income, education, insurance, and employment. Clinical characteristics of the cohort included: replacement steroid medications with dosage, reports of initial diagnosis events, satisfaction with current medication, adrenal crises and associated autoimmune and endocrine diseases. Common responses in text fields were grouped and quantified to identify most frequent responses in the following categories: “other autoimmune diseases”, “other endocrine conditions”, “other diagnoses prior to adrenal insufficiency diagnosis”, “causes of adrenal crisis”, “symptoms of adrenal crisis”, and “DHEA use”. The majority of data was from specific categorical responses built into the questionnaire. Percent values were calculated from denominator value of the cohort (n=541) except where a subset was specified.
Data was analyzed using JMP version 12 (SAS, Cary, NC, USA). Continuous variables were reported as means and standard deviations and significant differences were quantified using a two-tailed Student’s t test. Categorical variables were reported as number and percent with differences compared using the chi-square (χ2) and Fisher’s exact test. p-Values of less than 0.05 were considered significant.
The cohort showed a female predominance (83%) and was almost exclusively non-Hispanic White (n=537, 97%), with four African Americans, one Asian, and two American Indians. Further, five (1%) subjects reported Hispanic ethnicity. These results differ from the overall 2010 US census population which found 13% of the US population was African American, 5% Asian and 16% reporting Hispanic ethnicity. The majority of the study group was either working or retired (15% reported being disabled). Ninety-two percent reported education beyond high school, and health insurance was almost universally reported (Table 1). Income was generally high with only 12% of the cohort reporting annual income below $25,000/year, and 69% reporting an income of greater than $50,000/year.
Clinical characteristics were consistent with PAI patients of predominantly autoimmune origin (Table 2). Thyroid disease was a very common comorbid disease, with half of the cohort (n=270, 50%) reporting “any” thyroid problem and autoimmune thyroid disease explicitly reported by 216 (40%) subjects. Diabetes was the next most common endocrine association with 32 (6%) individuals responding, but there was no information regarding the subtype of diabetes (Type 1 or 2). Other autoimmune conditions reported included: vitiligo (n=38), autoimmune polyendocrine syndrome (APS) type 2 (n=35), celiac disease (n=25) and pernicious anemia (n=25).
More than half the cohort reported difficulty with initial diagnosis and provided additional comments about their experiences. On average, patients saw 3.5 physicians before they were diagnosed and there was a high rate of prior diagnoses that were focused on mental health – hypochondriasis, depression, and anxiety (18%) in patients who were later found to have PAI. These findings are buttressed by responses where 46% of patients reported being told that they were “overly concerned about health or did not have a physical problem”, and 30% reported they were advised to seek mental health counseling for their symptoms. In addition to the mental health diagnoses, gastrointestinal diagnoses were frequently noted – consistent with the common symptoms of anorexia, nausea, and diarrhea.
The majority of patients (n=434, 78%) of PAI patients were using hydrocortisone (HC), with prednisone (14.8%) and dexamethasone (1.8%) use was reported less frequently. The mean HC dosage reported was 24.4 mg/day, and the mean prednisone dose 7.3 mg/day. A mineralocorticoid replacement, fludrocortisone, was used by 426 (77%) of the subjects and dihydroepiandosterone (DHEA) usage was reported by 56 subjects (14%). Importantly, only 73% of the subjects felt that their overall regimen of steroid replacement was adequate for successful function in life.
Patient education about self-management of PAI varied (Table 3). A sizeable proportion (42%) of patients reported that they had not received instructions to use intra-muscular (IM) glucocorticoids for symptoms of acute crisis or injury, which is currently recommended in both US and European treatment guidelines , . Slightly more than 10% of patients said that they were unaware that they should modify their medications during situations of increased stress or illness.
Episodes of adrenal crisis were common, 61% of subjects reported one or more episodes requiring an emergency room visit since they were diagnosed (Table 4) and 11 noted that they were initially diagnosed during an adrenal crisis. In the emergency room, 35% reported delays getting the appropriate care; and half the respondents stated that the emergency room staff did not have a good understanding of the urgency to treat and/or what care was needed.
The precipitating events for an adrenal crisis in categorical responses were: stomach illnesses (n=165/341, 48%), cold or flu (104/341, 30.5%), surgery (47/341, 13.8%) and injury (29/341, 8.5%). Interestingly, 23% of the subjects described significant emotional stress such as family deaths as a crisis precipitant. Other triggers reported included infection, increased physical exertion, dehydration, heat, and inadequate steroid dose relative to the situation.
Symptoms associated with an impending crisis reported were: extreme weakness (98%), nausea (84%), and loss of consciousness (24%) in categorical responses. Other responses included vomiting and diarrhea, dizziness, generalized weakness, muscle cramps, headache, shaking/tremors, mental confusion, back pain, and cardiovascular symptoms. Factors associated with adrenal crisis are shown in Table 4.
Using data from a 2015 to 2016 US-based patient registry of more than 900 adrenal insufficiency patients and a case definition for PAI, we identified a group of 541 US patients that had characteristics similar to other published PAI cohorts. Issues common to previous reports were: (1) frequent comorbid autoimmune conditions including thyroid disease, diabetes, and vitiligo, (2) experience of diagnostic difficulty, (3) occurrence of adrenal crises, and (4) treatment inadequacy. Subjects reported typical medications/dosage regimens for PAI . These results suggest that the use of an online registry is a feasible method to assess the status of US patients with PAI. Although it is clearly desirable to confirm disease status with appropriate biochemical testing, there are challenges to recruiting a large adrenal cohort for clinical research using biochemical criteria.
Previous efforts to obtain information about US adrenal insufficiency patients have been limited and as a rare disease, it is difficult for a single clinical center to recruit sufficient subjects for study. Rare diseases have unique problems  but have been an area where patient registries have been able to play a role , , . Table 5 summarizes some of the potential sources of data for clinical research. Although evidence-based medicine efforts have focused on randomized clinical trials as providing the highest level of evidence for assessing treatments, generalizability to real world populations limits their usefulness and they have no obvious role in evaluating diagnosis accuracy. Rare diseases have previously been addressed with case series, but these commonly involve small numbers of subjects, often from a single center, and by virtue of being included in a case series – a diagnosis has been established with some degree of certitude. Administrative data can be a source of information about rare disease diagnostic error by identifying deviations from expected disease prevalences. Observational cohorts and the related patient registries can be broadly inclusive for rare diseases.
Several recent editorials address issues for patient registries in rare diseases ,  that recognize challenges with adequate diagnostic testing and inclusion of reliable information. However, strengths identified in registry data include: the ability to monitor the health of a disease population, a comprehensive view of patients and their characteristics, high rates of participation, and the value of longitudinal data for understanding the impact of treatments on outcomes. Patient registries and observational cohorts commonly include patient-reported data as well as biochemical tests and diagnostic procedures. Although extensive efforts have been made to obtain patient input in research and recognize the importance of patient-reported outcomes, there remains some doubt within medicine about the utility of patient input . However, patient input is essential to properly assess treatments, outcomes and especially diagnostic error.
In 1997, the National Adrenal Disease Foundation analyzed a mailed survey of 700 North American patients (http://www.nadf.us/articles/north-american-survey-of-individuals-with-addisons-disease-1997). Diagnostic difficulty for PAI patients in the US is a persistent problem. That 1997 survey found delays in diagnosis with 38% of subjects reporting delays of 1–5 years. Further, 43% were told their symptoms were likely psychological – before their ultimate diagnosis of adrenal insufficiency. In this 2015 study we found 57% of the patients reported difficulty with diagnosis, and a significant proportion (48%) were told that they did not have a physical problem. Although there may have been some overlap in participants in these two NADF surveys, the nearly 20-year interval and no increase in mean ages suggest they are distinct groups and that there has been no progress in reducing diagnostic problems.
Schiff et al. defined diagnostic error as “any mistake or failure in the diagnostic process leading to a misdiagnosis, a missed diagnosis or a delayed diagnosis” . The most common reason for error in their study was failure/delay in considering the potential diagnosis. A potential explanation for the difficulty in diagnosis of PAI may be related to a mismatch between the symptoms that physicians expect (afternoon fatigue, increased skin pigmentation, and hypotension) , and those that are commonly reported by patients (weakness, nausea, diarrhea, salt craving, muscle weakness, symptoms of unstable blood pressure, weight loss, anorexia, muscle cramps, mental fog/confusion) – such that the diagnosis is not considered . The differences between physician-expected symptoms from classic, late-stage descriptions and the broad range of actual patient-reported symptoms may both reduce diagnostic accuracy and feed the perception that these patients have mental health issues.
Disease management for PAI remains challenging with excess mortality and reduced quality of life noted in other cohorts . Sustained release glucocorticoid (GC) preparations have been studied in Europe but are not available in the US  and the available drug choices do not provide optimal coverage for diurnal variation in cortisol. Earlier publications had suggested decreasing doses of cortisol replacement would be helpful to reduce cardiovascular complications from excess GCs, but others have subsequently reported increased risk of adrenal crisis with inadequate replacement .
Our data on disease management in the US shows hydrocortisone in split doses is the most common choice of patients and physicians and matches with current guidelines . The mean daily dosage of hydrocortisone reported in our cohort was 24.4 (8.7) mg. In the 1997 NADF survey, the hydrocortisone equivalent dosage was approximately 30 mg, thus our findings are consistent with a trend toward lower replacement doses. However, 27% of our subjects reported that their overall steroid replacement was not adequate for successful function, where only 8% were dissatisfied in the 1997 NADF study, raising questions about the adequacy of disease management in the face of decreased replacement.
Overall, timely, accurate diagnosis, and adequate replacement therapies appear to be important problems in the US as in other countries, however, there has not been active research in the US to rigorously assess or address these issues. The cohort we studied was highly educated, relatively wealthy, and likely to advocate effectively for their health care. They experienced a range of symptoms across multiple body systems that were consistent with the disease, and yet nearly half were told that they did not have a physical problem and a third were referred for mental health counseling to deal with their symptoms. Based on the current data there has been no improvement in diagnostic accuracy since the 1997 survey.
As a first effort to assess adrenal insufficiency in the US, our study had some limitations. Selection of subjects was based on a case definition within a patient registry. Misclassification bias without biochemical studies is possible. As patients had provided data to the registry without knowing whether it would be used for research purposes we used medications and dose to supplement to the self-report of PAI to reduce the potential of falsely claiming the diagnosis. It is possible that some patients with SAI or unknown causes of AI were admixed. We would expect that associations to autoimmune conditions might be reduced with that admixture.
Patients who participate in an online registry may be different from those who do not participate and may express greater dissatisfaction with care. Patient registries are similar to observational cohorts and can provide important data in rare diseases where recruitment of adequate subjects is difficult or impossible . The population was an opportunistic sample – not population-based, but the optimal method for selecting a representative cohort for a rare disease within the US healthcare system may be difficult to define.
The demographic pattern favoring higher education and income has at least two possible interpretations: a recruitment bias, a real diagnostic disparity or some combination of those factors. Given the known difficulties with AI diagnosis and the existence of health disparities in other diseases – it seems appropriate to seek additional information to address that question. Future efforts to understand this might include enhanced recruitment and outreach to physician practices in underserved minority populations.
Future directions could include a more comprehensive patient registry in the US specifically designed for research and using standard methodology, with confirmation of specific diagnosis, longitudinal monitoring and further study of vulnerable populations. Adrenal insufficiency may be a sentinel diagnosis to spearhead processes for improving diagnostic error. It is a specific entity that could be used to develop tracking, educational initiatives and monitoring of disease prevalence as a marker of diagnostic success.
Bergthorsdottir R, Leonsson-Zachrisson M, Oden A, Johannsson G. Premature mortality in patients with Addison’s disease: a population-based study. J Clin Endocrinol Metab 2006;91:4849–53. CrossrefPubMedGoogle Scholar
Bensing S, Hulting AL, Husebye ES, Kampe O, Lovas K. Management of endocrine disease: epidemiology, quality of life and complications of primary adrenal insufficiency: a review. Eur J Endocrinol 2016;175:R107–16. Web of ScienceCrossrefPubMedGoogle Scholar
Hahner S, Spinnler C, Fassnacht M, Burger-Stritt S, Lang K, Milovanovic D, et al. High incidence of adrenal crisis in educated patients with chronic adrenal insufficiency: a prospective study. J Clin Endocrinol Metab 2015;100:407–16. CrossrefPubMedWeb of ScienceGoogle Scholar
Erichsen MM, Løvås K, Skinningsrud B, Wolff AB, Undlien DE, Svartberg J, et al. Clinical, immunological, and genetic features of autoimmune primary adrenal insufficiency: observations from a Norwegian registry. J Clin Endocrinol Metab 2009;94:4882–90. Web of ScienceCrossrefPubMedGoogle Scholar
Cooper GS, Bynum ML, Somers EC. Recent insights in the epidemiology of autoimmune diseases: improved prevalence estimates and understanding of clustering of diseases. J Autoimmun 2009;33:197–207. Web of SciencePubMedCrossrefGoogle Scholar
Meyer G, Badenhoop K, Linder R. Addison’s disease with polyglandular autoimmunity carries a more than 2.5-fold risk for adrenal crises: German health insurance data 2010–2013. Clin Endocrinol (Oxf) 2016;85:347–53. PubMedCrossrefGoogle Scholar
Bleicken B, Hahner S, Ventz M, Quinkler M. Delayed diagnosis of adrenal insufficiency is common: a cross-sectional study in 216 patients. Am J Med Sci 2010;339:525–31. CrossrefWeb of ScienceGoogle Scholar
Johannsson G, Falorni A, Skrtic S, Lennernäs H, Quinkler M, Monson JP, et al. Adrenal insufficiency: review of clinical outcomes with current glucocorticoid replacement therapy. Clin Endocrinol (Oxf) 2015;82:2–11. PubMedCrossrefGoogle Scholar
Bleicken B, Hahner S, Loeffler M, Ventz M, Decker O, Allolio B, et al. Influence of hydrocortisone dosage scheme on health-related quality of life in patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2010;72:297–304. CrossrefPubMedGoogle Scholar
Husebye ES, Allolio B, Arlt W, Badenhoop K, Bensing S, Betterle C, et al. Consensus statement on the diagnosis, treatment and follow-up of patients with primary adrenal insufficiency. J Intern Med 2014;275:104–15. CrossrefPubMedWeb of ScienceGoogle Scholar
Bornstein SR, Allolio B, Arlt W, Barthel A, Don-WauchopeA, Hammer GD, et al. Diagnosis and treatment of primary adrenal insufficiency: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2016;101: 364–89. Web of ScienceCrossrefPubMedGoogle Scholar
Kaufmann P, Pariser AR, Austin C. From scientific discovery to treatments for rare diseases – the view from the National Center for Advancing Translational Sciences – Office of Rare Diseases Research. Orphanet J Rare Dis 2018;13:196. PubMedCrossrefWeb of ScienceGoogle Scholar
Ambrosini A, Calabrese D, Avato FM, Catania F, Cavaletti G, Pera MC, et al. The Italian neuromuscular registry: a coordinated platform where patient organizations and clinicians collaborate for data collection and multiple usage. Orphanet J Rare Dis 2018;13:176. Web of ScienceCrossrefPubMedGoogle Scholar
Culver DA, Behr J, Belperio JA, Corte TJ, de Andrade JA, Flaherty KR, et al. Patient registries in idiopathic pulmonary fibrosis (IPF). Am J Respir Crit Care Med 2019. doi: 10.1164/rccm.201902-0431CI. [Epub ahead of print]. PubMedGoogle Scholar
Greenhalgh J, Dalkin S, Gooding K, Gibbons E, Wright J, Meads D, et al. In Functionality and feedback: a realist synthesis of the collation, interpretation and utilisation of patient-reported outcome measures data to improve patient care. NIHR Journals Library National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK, Southampton (UK), 2017. Google Scholar
Schiff GD, Hasan O, Kim S, Abrams R, Cosby K, Lambert BL, et al. Diagnostic error in medicine: analysis of 583 physician-reported errors. Arch Intern Med 2009;169:1881–7. PubMedCrossrefWeb of ScienceGoogle Scholar
Whittaker MJ, Debono M, Huatan H, Merke DP, Arlt W, Ross RJ. An oral multiparticulate, modified-release, hydrocortisone replacement therapy that provides physiological cortisone exposure. Clin Endocrinol (Oxf) 2014;80:554–61. CrossrefGoogle Scholar
Rushworth RL, Torpy DJ. Adrenal insufficiency in Australia: is it possible that the use of lower dose, short-acting glucocorticoids has increased the risk of adrenal crises? Horm Metab Res 2015;47:427–32. CrossrefWeb of SciencePubMedGoogle Scholar
About the article
Published Online: 2019-06-29
Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.
Research funding: None declared.
Employment or leadership: None declared.
Honorarium: None declared.
Competing interests: The funding organization(s) played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication.
Disclosure: The authors report no conflicts of interest in this work.