SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Objective

Women with rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) have fewer children on average than other women. We sought to determine the roles of infertility, pregnancy loss, and personal choice on family size in women with these diseases.

Methods

A reproductive history questionnaire was completed by women with RA and SLE participating in a longitudinal observational study. Within each disease cohort, participants were divided into 3 groups: those interested in having children at symptom onset who had either fewer children than planned (group A) or the same number as planned (group B), and those no longer interested in having children at diagnosis (group C).

Results

Of the 578 RA and 114 SLE women surveyed, >60% were in group C. Of those interested in having children, 55% with RA and 64% with SLE had fewer children than originally planned. Among women with RA, group A had 1 less pregnancy, 1 less live birth, and an infertility rate 1.5 times higher than group B; the miscarriage rate was similar in both groups. Compared to SLE group B, SLE group A had a similar number of pregnancies, but a 3-fold higher rate of miscarriage and 1 less live birth. Concerns about child health and personal welfare were associated with a lower pregnancy rate.

Conclusion

In this population, more than one-half of young women with RA or SLE had fewer biologic children than desired. While patient choice plays a role, infertility in RA patients and miscarriage in SLE patients are also important.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Women with chronic illnesses may have fewer children than planned, potentially affecting the pleasure and fulfillment these women are able to gain through family life (1). For example, a study of young women with cancer found that the one-third of the women who had fewer children than they had wanted reported significant emotional distress, particularly among the women who had no biologic children (2). As both rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) can strike women during their reproductive years, women with these diseases likely encounter challenges in family building.

Prior studies have demonstrated that women with RA have fewer children than do healthy controls, and suggested that this may be due either to decreased fertility or to personal choice (3–5). We sought to assess whether women diagnosed with RA prior to completion of childbearing had fewer children than those diagnosed later, and to determine the possible causes of such a difference, including infertility, pregnancy loss, and personal choice. Women with SLE, on the other hand, have been reported to have a similar number of pregnancies as other women, but a higher rate of pregnancy loss (6). We sought to determine if young women with SLE had fewer pregnancies and children than women diagnosed later in life, and the causes for these differences.

Significance & Innovations

  • More than one-half of women diagnosed with rheumatoid arthritis (RA) or systemic lupus erythematosus (SLE) prior to the completion of childbearing have fewer children than they hoped for.

  • Among women with RA, the rate of infertility is higher among women diagnosed during childbearing than those diagnosed once childbearing was complete. Women who had fewer children than desired had a high rate of infertility (42%).

  • Among women with SLE, infertility does not appear to be significantly increased, but having fewer children than desired was associated with pregnancy loss.

  • Concerns about the well-being of a pregnancy, offspring, and disability contributed to a decrease in the number of pregnancies conceived in women with RA and SLE.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

In the fall of 2009, a comprehensive reproductive health questionnaire was mailed to 1,017 female participants in the National Data Bank for Rheumatic Diseases (NDB) (7). Each woman had indicated that she would be interested in completing an additional questionnaire about her reproductive health history. Each woman had been diagnosed by her physician with either RA (n = 852) or SLE (n = 165) and had previously completed a recent NDB 6-month questionnaire. All responses were received within a 3-month time period. Demographic characteristics were previously collected by the NDB and included age, age at disease onset, race, education, and marital status. All respondents with RA or SLE were included in this analysis.

To determine each woman's reproductive history, the questionnaire first asked whether they had ever been pregnant, and then presented followup questions for further clarification. Women with prior pregnancies were asked to report the total number of pregnancies, the number of live births, the number of miscarriages, and the number of pregnancy terminations. Women were also asked to report whether they had ever been unable to become pregnant or been clinically diagnosed with infertility.

Once the reproductive background had been established, the questionnaire turned to issues of reproductive intent and the factors that could impact family building (Figure 1). Within each disease cohort, women were divided into 3 groups based on their responses: those who had fewer children than they had intended (group A), those who had the same number of children as they had intended (group B), and those who had completed childbearing at the time of diagnosis (group C).

thumbnail image

Figure 1. Questions posed to determine whether a woman chose to have fewer pregnancies because of her disease.

Download figure to PowerPoint

Infertility was ascertained using 2 questions, with a positive response to either question providing a diagnosis of infertility. The questions were: 1) Have you ever been unable to get pregnant after 12 months of trying? (yes/no) and 2) Have you ever been diagnosed with infertility by a physician? (yes/no). The followup question was: “If ‘yes,’ why were you infertile or unable to get pregnant?” The options for this were premature ovarian failure/premature menopause, problems with ovulation, problems with fallopian tubes, problems with uterus, endometriosis, elevated prolactin, unexplained infertility, male factor (problem with sperm), I don't know, and other. For the purposes of analysis, the first 2 options were combined and the description of “other” was reclassified, when possible. Finally, patients were asked if they had used physician-assisted reproductive methods to get pregnant and which types (oral or injectable medications, in vitro fertilization, or other).

Women with RA and SLE were assessed independently. The numbers of pregnancies, live births, and miscarriages were compared between groups A, B, and C for each disease. Chi-square testing was used to compare categorical data, and Student's t-test was used for comparing continuous data. Linear regression was used to assess the impact of independent predictors of pregnancy and live birth rates. To evaluate how the different reasons to avoid pregnancy affected family building, we analyzed the number of pregnancies for women in group A without infertility who answered yes or no to each concern. Finally, a multivariate regression analysis was used to determine how age at diagnosis, infertility, miscarriages, and the number of reasons given to avoid pregnancy influenced the numbers of pregnancies and live births. These factors were selected for their clinical relevance to the number of pregnancies and live births. Statistical analysis was done using Stata, version 10.1, with statistical significance set at a 2-tailed P value of 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

The response rate for the SLE cohort (69%, 114 responses from 165 questionnaires) was similar to the RA cohort (68%, 578 responses from 852 questionnaires; P = 0.76). The respondents were, on average, 3 years older than nonrespondents, with a smaller percentage of women who had never married (1.75% of respondents compared to 6.12% of nonrespondents; P < 0.01). Although respondents had a modestly longer disease duration (19.8 versus 18.1 years; P = 0.06), they reported a similar current level of global disease severity (P = 0.61), functional ability (P = 0.68), and overall quality of life (P = 0.67) compared to nonrespondents.

The respondents were primarily older white women (Table 1). Although the mean age at symptom onset was not significantly lower for women with SLE than for those with RA, more women with SLE were ages <35 years at the time of symptom onset (51.8%) than were women with RA (41.5%; P = 0.04). Women with SLE were also younger at the time they completed the questionnaire compared to women with RA. Although a small group of women with RA (4%) and SLE (6%) were ages <35 years at the time of questioning, and therefore perhaps had not completed childbearing, the majority of respondents were ages >35 years. A small minority, ranging from 1.6% with RA to 2.8% with SLE, had never been married.

Table 1. Demographics*
 RA (n = 578)SLE (n = 114)P
  • *

    Values are the number (percentage) unless otherwise indicated. RA = rheumatoid arthritis; SLE = systemic lupus erythematosus.

Current age, mean ± SD years58.6 ± 1253.8 ± 11< 0.01
Currently age <35 years23 (4)7 (6.1)0.3
Race   
 White537 (93.6)91 (84.3)< 0.01
 African American17 (3)5 (4.6) 
 Hispanic7 (1.2)8 (7.4) 
 Other13 (2.3)4 (3.8) 
Educational attainment, mean ± SD years14.6 ± 2.114.1 ± 2.60.02
Never married9 (1.6)3 (2.8)0.4
Age at symptom onset, mean ± SD years38.4 ± 14.735.7 ± 13.60.07
Group A: had fewer children than expected120 (20.8)29 (25.4)0.26
Group B: had the same number of children as expected95 (16.4)16 (14.0)0.52
Group C: had completed childbearing at the time of symptom onset363 (62.8)69 (60.5)0.50

The percentages of those reporting that they were still interested in having children at the time of symptom onset but had fewer children than expected (group A), that they were still interested in having children at the time of symptom onset and had the same number of children that they expected (group B), or that they had completed childbearing at the time of symptom onset (group C) were similar between respondents with RA and those with SLE. When taken in aggregate, the average numbers of pregnancies, children, and terminations were similar in both cohorts, but women with RA were more likely to report infertility and women with SLE had a higher rate of miscarriages (Table 2). In both cohorts, the majority of those who developed symptoms prior to the completion of family building ended up with fewer children than they had originally planned to have.

Table 2. Numbers of pregnancies, live births, miscarriages, and terminations, and rate of infertility among women with RA and SLE*
 RASLEP
  • *

    Values are the mean ± SD unless otherwise indicated. RA = rheumatoid arthritis; SLE = systemic lupus erythematosus.

No. of pregnancies2.9 ± 1.63.2 ± 1.70.18
No. of children2.5 ± 1.42.3 ± 1.30.34
No. of miscarriages0.5 ± 0.950.7 ± 1.20.04
No. of terminations0.15 ± 0.480.25 ± 0.520.06
No. reporting infertility (%)144 (25)18 (16)0.04

RA patients in groups A and B were, on average, 20 years younger at the time of symptom onset than were RA patients in group C (Table 3). At the time the questionnaire was completed, the women in group A were 7 years younger than the women in group B (P < 0.01), and 12% of the women in group A compared to 7% in group B (P = 0.2) were ages <35 years, the latter suggesting women who may not yet have completed childbearing. Among women with RA, those in group A had, on average, 1 less pregnancy and 1 less child than women in groups B or C. While the numbers of miscarriages and terminations were similar in all RA groups, women in group A reported a rate of infertility 1.5-fold higher than that reported in group B (P < 0.05) and 2.3-fold higher than that reported in group C (P < 0.05). The most common reason these patients gave for infertility was ovulatory dysfunction (19.4%), followed by unexplained infertility (15.3%) and endometriosis (10.4%). Of the infertile women, 52.0% in group A, 44.4% in group B, and 28.8% in group C tried infertility treatments. The majority used oral or injectable hormonal therapy, but 10 women reported trying in vitro fertilization.

Table 3. Pregnancy outcomes for women with rheumatoid arthritis
 Not finished having children at time of diagnosisFinished having children at time of diagnosis (group C)
Fewer children than desired (group A)Expected number of children (group B)
  • *

    P < 0.01 comparing women who had finished having children at the time of diagnosis with those who had not.

  • P < 0.01 comparing women who had fewer children than desired with those who had as many children as planned.

No. of women (%)120 (21)95 (16)363 (63)
Age at diagnosis, mean ± SD years26.9 ± 1125.8 ± 1345.6 ± 11*
Current age, mean ± SD years49.4 ± 1156.3 ± 1462.2 ± 10*
No. of pregnancies, mean ± SD2.2 ± 1.53.2 ± 1.73.1 ± 1.5*
No. of children, mean ± SD1.6 ± 1.02.7 ± 1.52.6 ± 1.3*
No. of miscarriages, mean ± SD0.5 ± 1.20.5 ± 1.00.5 ± 0.9
No. of terminations, mean ± SD0.21 ± 0.570.17 ± 0.510.13 ± 0.45
No. reporting infertility (%)50 (42)27 (28)65 (18)*

While infertility can account in part for why RA group A had fewer pregnancies than did group B, almost 60% of women in group A did not report infertility but still had significantly fewer pregnancies and children. This is not explained by pregnancy loss. It appears instead that this population of women chose to limit the number of pregnancies due to RA-related concerns (although, interestingly, physician recommendations to avoid pregnancy did not meaningfully affect the number of conceptions). Among fertile women in RA group A: 1) 53% reported concern about their ability to care for children, and among these women was a trend toward fewer pregnancies (2.1 per woman) than among women who did not report this concern (2.9 per woman; P = 0.1); 2) 37% expressed concern that their disease or medications would harm a baby, and those with this concern had fewer pregnancies (1.8 per woman) compared to those without it (2.9 per woman; P = 0.02); and 3) 17% were concerned that their children would develop RA, and among these women was a trend toward limiting the number of pregnancies (1.6 per woman) compared to those without this concern (2.6 per woman; P = 0.09). Reporting more reasons for not conceiving was also associated with fewer pregnancies. Fertile women with RA who listed no reasons for avoiding conception had, on average, 3.0 pregnancies, whereas those who listed 4 reasons had, on average, 1.5 pregnancies.

Women with SLE who had not completed childbearing at the time of symptom onset were on average 15 years younger at diagnosis than women who had (P < 0.001), but the current age was similar between the groups (Table 4). While the number of pregnancies, rate of infertility, and number of elective terminations were similar in all SLE groups, women in group A had, on average, 1 fewer child than did women in groups B or C. Overall, 21.7% of pregnancies conceived by women with SLE (either before or after the diagnosis) ended with a miscarriage. Although statistically nonsignificant (P = 0.16), women in group A had a 3-fold increase in the number of miscarriages relative to the other 2 groups. Women with SLE who reported disease-related child welfare concerns had fewer conceptions than women who did not indicate these concerns. Again, the pregnancy rate for women whose physicians advised against pregnancy was similar to that for women whose physicians did not.

Table 4. Pregnancy outcomes for women with systemic lupus erythematosus
 Not finished having children at time of diagnosisFinished having children at time of diagnosis (group C)
Fewer children than desired (group A)Expected number of children (group B)
  • *

    P < 0.01 comparing women who had finished having children at the time of diagnosis with those who had not.

  • P < 0.01 comparing women who had fewer children than desired with those who had as many children as planned.

No. of women (%)29 (25)16 (14)69 (61)
Age at diagnosis, mean ± SD years25.4 ± 1228.7 ± 1341.9 ± 11*
Current age, mean ± SD years52.0 ± 1150.4 ± 1555.4 ± 10
No. of pregnancies, mean ± SD2.9 ± 1.83.1 ± 1.03.3 ± 1.7
No. of children, mean ± SD1.6 ± 1.22.7 ± 0.82.6 ± 1.3*
No. of miscarriages, mean ± SD0.9 ± 1.60.3 ± 0.50.7 ± 1.2
No. of terminations, mean ± SD0.3 ± 0.60.2 ± 0.40.2 ± 0.5
No. reporting infertility (%)5 (17)4 (25)9 (13)

In linear regression models of women in RA groups A and B, older age at diagnosis was associated with a higher number of pregnancies, while infertility and reporting more reasons for avoiding conception were each associated with fewer pregnancies and fewer children (Table 5). In similar models for women with SLE, fewer pregnancies and fewer live births were associated only with the number of reasons given for avoiding conception; infertility, age of SLE onset, and miscarriage did not impact the number of pregnancies or live births.

Table 5. Multivariate regression models of factors affecting conception and family size among women who had not completed childbearing at symptom onset*
 RASLE
CoefficientPCoefficientP
  • *

    The factors included in this analysis were selected based on their clinical significance for these findings. RA = rheumatoid arthritis; SLE = systemic lupus erythematosus.

No. of pregnancies    
 Age at symptom onset, years0.020.020.0080.66
 Reported infertility−0.670.0050.740.18
 Reported number of reasons not to conceive−0.42< 0.001−0.370.02
No. of children    
 Age at symptom onset, years0.010.160.020.17
 Reported infertility−0.74< 0.001−0.030.94
 Reported number of reasons not to conceive−0.290.001−0.370.004
 No. of miscarriages−0.890.32−0.250.09

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

In this cohort of women with RA and SLE, more than one-half of those diagnosed prior to completing childbearing had fewer children than they had originally planned. Factors that appear to contribute to this pattern most substantially are infertility among women with RA, pregnancy loss in women with SLE, and in both groups, patient concerns about the impact of rheumatic disease on offspring and family life.

While the unique self-selected population included in this study may limit the generalizability of these results, the similarities in our findings to prior studies suggest that the findings of increased infertility in women with RA and pregnancy loss in women with SLE are sound. The response rate for this study was 68–69%, perhaps lower than would be expected. It is possible that more women with difficult reproductive histories would have the motivation to complete this long survey, thus skewing the data toward a higher rate of infertility and/or pregnancy loss.

The rates of pregnancies and complications in this study, however, are similar to what have been reported in previous studies of women with RA and SLE. The average number of pregnancies and miscarriages per woman with RA (2.9) is similar to that reported by Katz (2.7) (8). A recent comparison of pregnant women with and without RA in Denmark found a higher rate of fertility treatment (9.8% versus 7.6%) and requiring >12 months to conceive (25.0% versus 15.6%) among women with RA (5). The overall rate of miscarriage in the patients with SLE in this study (21.7%) matches that found in retrospective cohort studies by both Petri and Allbritton in Baltimore (21.2%) and Hardy et al in the UK (22.1%) (6, 9). Overall, the number of children reported for women with RA (2.5) and SLE (2.3) in this study is similar to what would be expected in the general population over the past 50 years in the US, when the total fertility rate has ranged between 1.80 and 2.07, meaning that the average woman would have 2 children over her lifetime (10).

As early as 1958, studies reported that women with RA had fewer children than their peers. In addition, women diagnosed with RA before age 18 years or before bearing children were found to have fewer total children than women diagnosed later in their lives (3, 4, 11). Other studies detected no differences in conception rates, but these were early investigations that did not distinguish between a woman's intention to conceive and her ability to conceive (12, 13).

While approximately 10% of the general population today reports infertility, the rate of infertility among RA patients enrolled in the current study was 25%, ranging from 18% for women who had completed childbearing prior to diagnosis to 42% for women who had fewer children than planned. Women with SLE, on the other hand, do not appear to have an increased rate of infertility. Why infertility disproportionately affects women with RA relative to women with SLE remains unclear, because on the surface, the 2 patient populations share primary vulnerabilities (as both diseases are characterized by systemic inflammation), affect women during their reproductive years, and require similar medications, but several theories may warrant consideration, including RA-induced alterations in cytokines (in particular interleukin-6 or tumor necrosis factor) or regulatory T cells (14–16). Medications, in particular nonsteroidal antiinflammatory drugs, may also play a role in infertility (17). Endometriosis was only reported in 10% of women with infertility, less than the rate expected in the general infertile population (25–50%) (18). While some prior studies have suggested an immunologic link between endometriosis and rheumatic disease, this is not supported by this study (19, 20). Several recent larger studies have also failed to identify a clear link between these processes (21–23).

That nearly one-third of women diagnosed with RA prior to completion of childbearing experience infertility demands further investigation, with an eye on the impact that current medications have on fertility, both negative and positive. We found that concerns about how the disease might affect child welfare appear to explain the lower number of pregnancies among those not affected by infertility.

Although women with SLE have normal rates of fertility, prior studies have demonstrated that pregnancy loss occurs more frequently than normal in this population (6, 24). Retrospective cohort studies demonstrate that the miscarriage rate prior to diagnosis of SLE is also higher than among healthy controls, and that this rate is higher still for women diagnosed with SLE prior to pregnancy (6). In prospective cohorts, the rate of miscarriage is similar between women with SLE and healthy women, but the rate of stillbirth is 5- to 10-fold higher in patients with SLE than in the general population (24–26). Pregnancy loss in SLE is most strongly associated with increased SLE activity prior to and during pregnancy (25). It may be hypothesized that systemic inflammation from SLE activity impairs placentation and promotes cervical ripening or preterm labor, all of which increase the risk for pregnancy loss. Further investigation into the biologic causes of pregnancy loss in women with SLE may lead to advances in care during this period and improve the pregnancy success rate.

Our study suggests another factor that plays a significant role in determining family size: personal choice. That is, a substantial percentage of the reproductive-age participants expressed concern for themselves and their potential offspring, primarily with regard to the practical and physical limitations imposed by disease, the side effects of medication, and genetic risks. When such concerns were present the number of conceptions dropped, and if multiple concerns were reported, the trend was even more pronounced.

Concerns about the adverse effects that RA or SLE disease activity might have on the health and care-giving capacities of the mother and therefore on the well-being of the child led to a significant decrease in the number of pregnancies. This dynamic appears to explain the unusually low number of pregnancies among women with RA and SLE who do not report infertility. While some of these concerns are well-founded (chronic pain, stiffness, and fatigue, for example, can make caring for an infant and child difficult), improved disease control through the use of new and more aggressive immunosuppressants may lessen the degree of concern in the coming years.

Another key misgiving expressed by the study population centered on the potential deleterious effects of drug therapy on offspring. Some medications, in particular, methotrexate, leflunomide, cyclophosphamide, and mycophenolate mofetil, are known or suspected teratogens and should be avoided prior to and during pregnancy (27). Others, including hydroxychloroquine, azathioprine, sulfasalazine, and corticosteroids, pose little risk to a developing fetus (27). It is essential to inform patients with rheumatic disease about their pharmaceutical options and the risks associated with each option so that they can properly assess the risks of pregnancy in the context of RA or SLE and make optimal decisions for their own health and the health of their children. Indeed, in another section of this same survey, participants actually underestimated the teratogenic risks of many of these medications (28).

Women in this study also reported a high degree of concern about passing their rheumatic disease on to their offspring. In fact, the likelihood that their children will develop either SLE or RA is low. A child whose mother has RA has a 3-fold increased risk of developing that condition sometime in life, meaning that an estimated 3% of the offspring of women with RA will develop it themselves. However, these offspring are also more likely to develop other rheumatic diseases at any time in life, including a 1.5–2.5-fold increased risk for SLE, ankylosing spondylitis, scleroderma, Sjögren's syndrome, or hypothyroidism (29).

Our study has several limitations inherent to relying on patients' voluntary participation with mailed questionnaires. While we describe the important differences between those that completed the questionnaire and those that did not, it is likely that those who were invited are more educated and more likely to be white than the general US RA and SLE population. In addition, because many of the survey questions involved sensitive topics, it could be that there was a response bias toward women who were comfortable answering these questions. While it is unknown if any of the factors associated with participation in the NDB or in this survey influenced the outcomes we evaluated, it is difficult to say how our findings might generalize to the population of all women with RA or SLE. Nevertheless, the relatively large sample size provides useful results for future work.

This study highlights several important reproductive health concerns for women with RA and SLE. More than one-half of women diagnosed with RA or SLE prior to the completion of childbearing have fewer children than they had hoped for. In both groups of women, concerns about inability to care for a child, damage from medications, and genetic transmission of their disease led to fewer pregnancies. Concerns about disability impairing the ability to care for children can only be addressed through improved disease management. Concerns about medication effects on offspring, however, should be managed through improved education of physicians and patients; there are safe medical options during pregnancy that can both control disease activity and prevent negative outcomes for the developing fetus. Similarly, concerns about genetic transfer of rheumatic disease may be exaggerated and potentially reversed with improved patient education. Beyond patient choice, however, infertility in women with RA and pregnancy loss in women with SLE resulted in fewer children. Therefore, to help women fulfill their childbearing goals, we must address patient education as well as the underlying causes of infertility and pregnancy loss.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Michaud had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Clowse, Chakravarty, Costenbader, Chambers, Michaud.

Acquisition of data. Clowse, Chakravarty, Michaud.

Analysis and interpretation of data. Clowse, Chakravarty, Costenbader, Chambers, Michaud.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES