Only as strong as the weakest link: Adrenal insufficiency in the COVID-19 storm

During the last 3 years, the world has faced an unprecedented and unexpected challenge as the SARS-CoV-2 pandemic unfolded. Despite best efforts from healthcare professionals, governments and organizations, many people have died, more have been hospitalized and even more have suffered consequences from SARS-CoV-2 infection, in the short or long term.

In this issue of the Journal of Internal Medicine, Bergthorsdottir et al. have analysed the COVID-19-related outcomes in patients with adrenal insufficiency (AI) in Sweden [1]. Through a nationwide project that involved the entire Swedish population in the development of a unified healthcare database, they designed a matched-control study of mortality and morbidity associated with SARS-CoV-2 infection that yielded concerning results. With 5430 enrolled patients and a 10:1 ratio of sex- and age-matched controls, the authors have provided a relevant sample size for their investigations, gaining the leverage to answer a question that most cross-sectional studies could not: How did the COVID-19 pandemic impact patients with AI compared to the general population?

The primary study outcomes were the rate of positive SARS-CoV-2 tests, hospitalization, intensive care admission and death due to COVID-19. Even after adjusting for socio-economic factors, patients with AI had more than twice the risk for hospitalization, intensive care admission and death compared to non-AI controls, with an overall mortality rate of 0.8% compared to 0.2% in the matching cohort.

Interestingly, the authors have not found a significant difference in the positive testing rate and confirmed diagnosis of SARS-CoV2 infection. This seems to suggest that patients with AI are not increasingly susceptible to SARS-CoV2 infection but, rather, more prone to developing severe COVID-19. However, as Bergthorsdottir et al. argued, patients might have been more proactive in establishing containment measures such as isolating, hand washing or mask-wearing compared to the general population. This should have rather resulted in a decreased rate of infection, but the study is not powered to make further assumptions on this topic. Moreover, patients with AI and their healthcare providers could be more liable to search advanced medical care for SARS-CoV2 infection, but results on post-hospitalization outcomes seem to rule out this aspect as a potential confounder.

Aside from the results on prevalence, mortality and severity of disease on the overall population, the authors also ran further analyses that shed more light on a large AI cohort. Surprisingly, sex stratification did not result in differences in terms of severity or outcomes of COVID-19 disease. Early research [2] highlighted an increased severity of the SARS-CoV2 infection and related disease in the male population. Several speculations have been made, mostly pointing towards an increased inflammatory response in males firing an uncontrolled immune cascade that led to worse outcomes [3, 4]. The sex dimorphism seems to disappear in the AI cohort, suggesting that hypocortisolism, its treatment or both might shift patients’ immune profiling towards a more ‘inflammatory-prone’ phenotype, similar – and worse – to that observed in the more vulnerable male population. Moreover, a potential interference of glucocorticoids on physiological sex hormone secretion could have a further impact.

This is consistent with previous research showing an increase in inflammatory monocytes and a decreased expression of CD16 (the FYIII-Receptor) on natural killer cells in AI [5]. Chronic low-grade inflammation that seemed to be sustained by non-physiological replacement [6] could have contributed to the increased risk for worse COVID-19-related outcomes.

Moreover, decreased endogenous cortisol secretion could fail to control the cytokine storm following SARS-CoV2 infection in the lower respiratory tract.

Second, subgroup analyses for AI aetiology showed that patients with secondary AI had an increased frequency of hospitalization, intensive care unit admission and death compared to patients with primary AI. However, patients with SAI are older and have other hormone deficiencies, probably contributing to these results.

The frequency of diabetes, hypertension and chronic lower respiratory disease was also higher in patients with AI than in the age- and sex-matched controls. This is consistent with emerging data on increased risk for metabolic and cardiovascular diseases in chronic AI, but after adjustment for comorbidities, the hazard ratios in patients with AI remained increased for all COVID-19-related outcomes (hospitalization, intensive care unit admission and death) compared to controls.

One of the main limitations of the study is the lack of data on replacement therapies and adjustments during inpatient management. Most patients were taking oral hydrocortisone, but identification of different hydrocortisone formulations, dosing or regimen was not available. Adequate replacement, physiological therapies and prompt administration of stress doses in case of necessity are the cornerstone of AI management, and deviations from correct protocols [7] could have had an impact on mortality and morbidity during the COVID-19 pandemic. However, the study is a precise representation of outcomes from current standard-of-care in a country in which awareness and knowledge of AI are generally high. Despite the missed opportunity to stratify outcomes for treatments, doses or other clinical indexes (such as BMI or clinical presentation), we must acknowledge the value of unbiased results.

Lastly, due to the selected study time frame, the trial does not provide strong results on the effectiveness of vaccination in the AI population. More studies will be required to investigate the response to controlled antigen exposure in AI and the ability of SARS-CoV2 vaccines to mitigate the disease in this condition.

All in all, this clinical trial has confirmed an encompassing dysregulation of immune response in AI. Due to its characteristics, COVID-19 has hit patients with increased inflammatory response more severely. Based on previous studies, we can infer that patients with AI fall into this category. Despite all efforts to sufficiently replace endogenous cortisol secretion, achieve electrolyte balance, normalize arterial pressure and all the other outcomes that are routinely evaluated in chronic adrenal failure, the immune system remains a ‘weak link’ in the chain.

Sometimes, scientific research provides results that seemingly lack a strong clinical impact on general healthcare. The COVID-19 pandemic has given a tragically loaded answer to most of the ‘why do we care?’ arguments on fine-grain research in rare conditions such as AI. Instead, striving to achieve more than ‘good’ control of diseases is exactly what is expected from current research in this field. The reason to care, as the article from Bergthorsdottir et al. suggests, is as simple as it is important: In case the storm hits again, we must leave no one behind.

The author declares no conflicts of interest.