Carriers of a Classic CYP21A2 Mutation Have Reduced Mortality: A Population-Based National Cohort Study

Abstract

Context

Congenital adrenal hyperplasia (CAH) is a common monogenic recessive disorder. It has been suggested that CYP21A2 deficiency is common because carriers may have a survival advantage, 1 in 15,000 in most populations. Carriers of CYP21A2 mutations typically do not have clinical symptoms but have a defined phenotype with a more prompt cortisol response to ACTH.

Objective

We investigated whether the mortality was lower, and determined the cause of death in carriers and population controls.

Design

A total of 1143 obligate carriers of a CYP21A2 mutation (561 men) were identified from the Swedish National CAH Registry, encompassing >700 patients and the Multi-Generation Registry to identify their parents. The mortality and cause of death were identified through the Swedish Cause of Death Registry. The hazard ratios (HRs) and 95% CIs were calculated. The results were compared with controls from the general population, matched for sex and age.

Results

The overall mortality was lower in carriers of a CYP21A2 mutation compared with the controls (HR 0.79; 95% CI, 0.678 to 0.917; P = 0.002). The difference was more marked among carriers of a more severe mutation. Infection as the cause of death was significantly lower (HR 0.65; 95% CI, 0.48 to 0.87; P < 0.01), particularly for death in pneumonia (HR 0.22; 95% CI, 0.06 to 0.88; P = 0.03). The lower overall mortality among women compared with men in the general population was confirmed among both carriers and controls.

Conclusion

Obligate CYP21A2 carriers of a classic mutation had a reduced mortality. Specifically, a possible reduced mortality due to pneumonia was seen.

Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency has an incidence of 1 in 10,000 to 20,000 in most populations. It is one of the most common monogenic autosomal recessive disorders (1). The incidence is even higher in some populations, for example, 1 in 295 among Yupik Eskimos (classic CAH) and 1 in 27 in the Ashkenazi Jewish population (nonclassic CAH) (2). The carrier frequency in the Swedish population is about 1 in 50 people for a mutation in the CYP21A2 gene and 1 in 70 for a classic CYP21A2 mutation (3).

It has been suggested that CYP21A2 deficiency is common because it may confer a survival advantage to carriers. The hypothesis is that it may be analogous to sickle cell anemia, whose carriers a more resistant to malaria. A survival advantage for carriers would explain the high frequency of CAH in most populations worldwide (4–6).

CAH due to 21-hydroxylase deficiency results in varying degrees of cortisol deficiency and concomitant increased androgen synthesis driven by the increased ACTH production. The severity of disease correlates with the severity of the CYP21A2 mutations (7, 8). CAH is clinically classified into the salt-wasting (SW), simple virilizing (SV), and nonclassic congenital adrenal hyperplasia (NCAH) (1). If untreated, patients with the SW form develop lethal salt crises in the neonatal period, and children with classic CAH (SW and SV) and 46,XX karyotype are born with virilized external genitalia. In contrast, patients with NCAH often do not come to medical attention until they develop signs of androgen excess over time, such as hirsutism or infertility, and men with NCAH often are not diagnosed (9, 10).

Carriers of CYP21A2 mutations do not have clinical symptoms, but they have a defined phenotype. The hypothalamic-pituitary-adrenal (HPA) axis is functionally different in that ACTH stimulation results in increased 17-hydroxyprogesterone levels, the metabolite before the enzyme block (11). In addition, the cortisol response to ACTH stimulation has been shown to be both more prompt and greater in carriers than in healthy, noncarrier controls (4, 5). This may result in a faster and more prompt mobilization during somatic stress. This increased HPA axis response may be caused by upregulated ACTH activity as a result of the relative deficiency in cortisol production (and feedback inhibition) because it has been shown that CYP21A2 mutation carriers have a lower 24-hour urinary excretion of free cortisol (5). It has been hypothesized that this prompt cortisol response could also enable a more rapid return to homeostasis and hence be advantageous (4). Thus, the increased capacity to synthesize cortisol in acute situations could be one explanation for an evolutionary advantage of being a carrier of a CYP21A2 mutation.

In a previous study, we compared the risk of receiving a psychiatric diagnosis in carriers of a CYP21A2 mutation by comparing parents of a child with CAH with parents of a child with hypospadias, parents of a child with diabetes, and the general population (12). We found that parents of children with CAH were less likely to have a psychiatric diagnosis before the birth of the child compared with all the other parents. In addition, after the birth of the affected child the difference between the parents of a child with CAH was even larger than that of the parents of children affected with hypospadias or diabetes. Our interpretation was that the parents of children with CAH (i.e., the CYP21A2 carriers) were less vulnerable to psychological stress. Interestingly, our findings are in contrast with the previous suggestion that CYP21A2 carriers are at higher risk of affective and anxiety disorders (5).

Epidemiological data show that girls and women have overall lower mortality than boys and men. Girls survive infections and trauma better and have lower mortality in the neonatal period than boys, and the mean age at death is lower in men than in women (13). It has been known for several decades that girls and women have a higher immune capacity, higher immune globulin levels, and a stronger humoral and cell-mediated immune response than men and boys (14). These traits can provide protection against infections but may lead to more inflammatory response and autoimmunity. It has been hypothesized that this difference in immunity between men and women is at least partly due to genetic differences (i.e., affected by genes located on the sex chromosomes) and due to the different levels of sex hormones men and women are exposed to. A number of immune-related regulatory genes are located on the X chromosome, such as the gene for the Toll-like receptor 7 and 8 (viral response) FOXP3 gene (a transcription factor regulating T-cells and CD 40 ligand) (15).

Androgens and estrogens have different effects on immune responses and affect different cell types (13). Estrogens have activating effects on humoral and cell-mediated immune responses, whereas androgens may result in a less efficient response (13). Carriers of CYP21A2 mutations have an efficient cortisol response but produce more adrenal metabolites with androgen effects. A negative effect on the immune response would therefore be possible.

The aims of this study were to investigate the mortality and the cause of death in obligate CYP21A2 mutation carriers (i.e., parents of children with CAH). We wanted to investigate whether the carrier phenotype could confer a survival advantage also with respect to stressful somatic events. In particular, we wanted to investigate whether mortality was affected in the carriers. Infectious diseases represent a large cause of death, especially from the evolutionary perspective of the history of humankind, and a more effective stress response or resistance to infections could therefore confer a survival advantage explaining the high percentages of CYP21A2 carriers in most populations. Because the overall mortality and the immune response differ between men and women, we used controls from the general population matched for sex and year of birth.

Methods

Subjects

Obligate carriers of CYP21A2 mutations were identified through the Multigeneration Register as parents of patients with CAH. The patients identified were born between 1910 and 2013, registered in the Swedish National CAH Registry or identified through the National Patient Register. The procedure has been described in detail previously (16–18).

A matched control design was used. Descriptive statistics are shown in Tables 1 and 2. In total 714 patients were identified through the Swedish National CAH Registry (640 patients) and through the National Patient Register, where an additional 74 patients were identified via the International Classification of Diseases, ICD-8 (255.01, 255.08), ICD-9 (2552, 255C), and ICD-10 (E25.0). Patients who were registered three or more times were included if no alternative diagnoses registered for that particular patient were more likely to be the final diagnosis. The CYP21A2 genotype was known for >80% of the patients in the National CAH Register.

The obligate carriers of CYP21A2 mutations (i.e., the parents of patients with CAH) were divided into groups based on the genotype or the clinical severity of CAH of their child. The parents of patients with SW or SV CAH were assumed to carry a classic CYP21A2 mutation. The other group consisted of parents of patients with NCAH, who may carry either a classic mutation or a mutation that would predict NCAH, in addition to the parents of patients with CAH of unknown severity. A total of 1143 carriers were identified (561 men and 582 women). They were divided into two groups comprising 731 carriers of classic mutations (359 men and 372 women) and 412 carriers of a mixture of mutations and of undetermined severity (202 men and 210 women). One hundred controls per patient were randomly selected from the National Population Register and matched for sex and age. The study population is described in Table 1.

Study protocol

The mortality and causes of death were investigated via the Swedish Cause of Death Registry. The cause of death was grouped according to the following ICD codes cancer (ICD-8 codes 140 to 239, ICD-9 codes 140 to 239, ICD-10 codes C00 to D48), diabetes (ICD-8 code 250, ICD-9 code 250, ICD-10 codes E10 to E14), adrenal (ICD-8 code 255, ICD-9 code 255, ICD-10 codes E25 and E27), dementia (ICD-8 code 290, ICD-9 code 290, ICD-10 codes F00 to F03 and G30), cardiovascular disease (ICD-8 codes 390 to 453, ICD-9 codes 390 to 453, ICD-10 codes I00 to I82), stroke (ICD-8 codes 430 to 436, ICD-9 codes 430 to 436, ICD-10 codes I60 to I64), ischemic heart disease (ICD-8 codes 410 to 414, ICD-9 codes 410 to 414, ICD-10 codes I20 to I25), chronic obstructive pulmonary disease (ICD-8 codes 491 to 492, ICD-9 codes 491 to 492, ICD-10 codes J43 to J44), and infectious diseases (ICD-8 codes 000 to 136 and 460 to 486; ICD-9 codes 001 to 139, 460 to 466, and 480 to 487; ICD-10 codes A00 to B99 and J00 to J22).

The study was approved by the Regional Ethical Review Board in Stockholm, Sweden.

Statistical analysis

The association between mortality and cause of death was assessed via relative risk analysis. Results were reported as hazard ratios (HRs) and 95% CIs. A P <0.05 was considered significant. Fisher exact test was used to verify the outcome in analyses of subgroups of less than five subjects.

Results

The overall mortality was lower in carriers of a CYP21A2 mutation compared with the general population (HR 0.79; 95% CI, 0.68 to 0.92; P = 0.002); the difference was significant also when men and women carriers were assessed separately compared with the controls (P = 0.02 and P = 0.044, respectively). For carriers of a classic mutation the HR was 0.68 (95% CI, 0.558 to 0.832; P = 0.0002). The group with a more severe mutation seemed to have a slightly lower risk (HR 0.65; 95% CI, 0.489 to 0.865; P = 0.003) than the SV group (HR 0.71; 95% CI, 0.538 to 0.943; P = 0.02). In all groups the women had a lower HR than the men; in other words, the lower overall mortality seen in the general population among women was confirmed in our study, among both carriers and controls (P < 0.001). The results are presented in Table 2.

Parents of children with NCAH or CAH of unknown severity (i.e., carriers of both severe and milder CYP21A2 mutations) did not show significant differences compared with controls in any of the parameters studied, indicating that the severity of the mutation may be important.

The different causes of death retrieved from the Cause of Death Registry are presented in Table 3. Infection as the cause of death was significantly lower among the carriers than controls (HR 0.65; 95% CI, 0.48 to 0.87; P < 0.01). In particular, a tendency for lower mortality in pneumonia was seen (HR 0.22; 95% CI, 0.06 to 0.88; P = 0.03, and Fisher exact test P = 0.056) (Table 3). The difference was not significant for sepsis. There were no reported deaths for other major infections such as erysipelas, influenza, or viral hepatitis. No difference in the overall mortality in cancers, ischemic heart disease, or cardiovascular disease was detected (Table 3).

Discussion

This is the largest study investigating carriers of a CYP21A2 mutation and the only one studying mortality. The study population was derived from a population-based national cohort. The obligate carriers (i.e., parents of patients with known CAH) had lower mortality. We found indications that they had a lower risk of death due to infection, specifically pneumonia. This lower mortality was more pronounced for carriers of severe (i.e., classic) CYP21A2 mutations and was not seen for the group with unknown or milder mutations. We did not see lower mortality in cardiovascular disease, as may have been anticipated considering the antiatherogenic profile reported by Livadas et al. (6).

It is not possible to determine the exact mechanism in this type of epidemiological study, but it can be speculated that being a carrier is associated with adaptive changes in the HPA axis rather than differences in the immune system. Carriers of severe CYP21A2 mutations have a prompt and more elevated increase in cortisol in response to ACTH. An acute cortisol response is important in all events of somatic stress (19) and may be particularly important during serious infectious. The timing of the glucocorticoid effect may be crucial, not only for an urgent response but also to prevent the effects of cytokines and inflammatory factors that can lead to glucocorticoid resistance and therefore prevent an effective glucocorticoid later during the disease process. It has been shown in mice that a lethal dose of TNF-α inhibits the glucocorticoid receptor function and induces glucocorticoid resistance, and that this can be prevented if dexamethasone is given before but not after the TNF-α. Hence, earlier treatment with glucocorticoid prevents the lethal effect of the TNF-α (20).

Treatment with glucocorticoids in combination with antibiotics for bacterial meningitis has been used since the 1950s. The timing of the glucocorticoid is considered important: the instruction is to initiate the adjuvant treatment with dexamethasone before or together with the first dose of antibiotics and continue for 2 to 4 days. The treatment is effective in preventing hearing loss in high-income countries, but the effects on mortality are contradictory and still under discussion (21). The timing is difficult to assess, and it is unclear whether it is the glucocorticoid treatment or the initiation of antibiotics that is relevant. In a study of 1746 patients, of whom 989 were treated with glucocorticoids within 1 hour of initiation of antibiotics, mortality was significantly lower in the glucocorticoid-treated group (8.9%) than in the group not treated with glucocorticoids (17.9%) (odds ratio 0.57; 95% CI, 0.40 to 0.81) (22). However, the evidence for the effects of glucocorticoid treatment on sepsis mortality is still controversial (23, 24).

Only two of the carriers in our study had pneumonia as the cause of death. It can therefore be speculated that a contributing factor to the lower mortality in infections, especially pneumonia, is the prompt and exaggerated cortisol response to stress in CYP21A2 carriers that may enable an early enough increase in cortisol levels to be effective. The more prompt and elevated cortisol response to ACTH may enable inhibition of inflammatory mediators and thereby counteract the development of glucocorticoid resistance via inhibition of transactivation and synthesis of glucocorticoid receptors seen after stimulation with inflammatory cytokines such as TNF-α and IL-6. The Cochrane Report from 2017 on pneumonia states that corticosteroid therapy was associated with lower all-cause mortality and lower mortality for severe community-acquired pneumonia (25). In addition, a recent meta-analysis concludes that adjunct glucocorticoid treatment had favorable effects on severe pneumonia (26). This speaks in favor of the faster cortisol response to stress as a possible explanation for the lower mortality in CYP21A2 carriers.

Carriers of CYPA2 mutations slightly higher androgen production. In a study by Charmandari et al. (5), CRH stimulation resulted in an area under the curve for ACTH that was larger in carriers than in controls. This difference did not result in a difference in cortisol response, but the carriers had a larger area under the curve for 17-hydroxyprogesterone (i.e., higher androgen levels). However, if the higher androgen production among the carriers, although in most cases a modest increase, would have a measurable effect on immune response, it would probably lead to a negative effect on the immune response rather than a more efficient defense against infections. Androgen effects on the immune system are not well studied but have been reported to reduce natural killer cell activity, downregulate NF-κB, and increase anti-inflammatory cytokines (14) and to have negative effects on the activation of a number of immune factors (15). In addition, if the effect seen in our study was mediated by a more efficient immune system in general, one could speculate that other diseases related to immune function would be affected, such as cancer. In this study, we did not see any differences in mortality related to cancer between the CYP21A2 carriers and the population controls, speaking against a hypothesis that the carriers would have more efficient or upregulated immune system in general.

Our previous study on psychological vulnerability in carriers of CYP21A2 showed a better ability to cope with psychological stress among heterozygous carriers of CYP21A2 mutations compared with parents from the general population, parents of children with hypospadias, and parents of children with diabetes mellitus type 1 (12). This ability may also contribute to a survival advantage and the high incidence of CYP21A2 carriers seen worldwide.

In this epidemiological registry study, only associations can be identified. Causal factors for the lower mortality in infectious diseases cannot be deciphered in detail, and investigation and follow-up of the results are not possible by virtue of the study design. It could be speculated that the parents of patients with CAH have a lower threshold for seeking medical attention, but they will not be treated by the child’s doctor. This difference may not have a large impact on mortality; an effect on treatment frequency would be more likely. Obesity was not studied here, although it is known to be a contributing factor in all-cause mortality (27).

The strengths of the study are that it covers a population-based cohort, implying no selection bias. The National CAH Registry covers almost all patients in Sweden, and the National Patient Register has had full coverage since 1964 for hospital care and since 2001 for outpatient care. The CAH diagnosis was validated in the registry, which was the basis for this anonymized epidemiological registry study, and thus it is unlikely that patients were misdiagnosed. The multigenerational registry has full coverage for the study population. Less than 1% of patients with CAH have a spontaneously occurring mutation (i.e., the likelihood that one of the parents is not a carrier is low and unlikely to affect our results). The carriers were compared with control groups from the population register, matched for sex and year of birth. The Cause of Death Registry has full coverage and good accuracy for the diagnoses.

Conclusion

This study found that carriers of CYP21A2 mutations may have a lower mortality from severe infections, possibly pneumonia in particular. This difference may have contributed during evolution to a survival advantage and the high incidence of CYP21A2 carriers seen worldwide.

Carriers have a more efficient cortisol response and produce more adrenal metabolites with androgen effects. The androgen excess is not a likely explanation for the observed difference because androgens have negative effects on the immune system. The prompt and elevated cortisol response to ACTH seems to be the more plausible explanation, enabling an acute and efficient stress response. This possible advantage with respect to somatic stress should be investigated further.

Acknowledgments

Financial Support: This project was supported by grants from the Magnus Bergvall Foundation (grant no. 2017-02138 and 2018-02566, to A. Nordenskjöld), Karolinska Institutet (to H.F.), the Stockholm County Council (to A. Nordenström), and the Swedish Research Council (no. 2017-02051, to C.A.M.) and via the Swedish Initiative for Research on Microdata in the Social and Medical Sciences (SIMSAM) Framework Grant no. 340-2013-5867 (to C.A.).

Author Contributions: A. Nordenström designed the study, collected the data, and wrote the manuscript. A. N., C.H., and H.F. analyzed the data. C.N., A. Nordenskjöld, L.F., and H.F. contributed in the study design and data collection. A. Nordenström, J.S., S.L., L.F., C.N., C.A., A. Nordenskjöld, and H.F. interpreted the data and critically revised the manuscript.

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.

Abbreviations:

Abbreviations:
 
• CAH

  congenital adrenal hyperplasia

 
• HPA

  hypothalamic-pituitary-adrenal

 
• HR

  hazard ratio

 
• ICD

  International Classification of Diseases

 
• NCAH

  nonclassic congenital adrenal hyperplasia

 
• SV

  simple virilizing

 
• SW

  salt wasting

References and Notes