Improved understanding of sex and gender-specific differences in the aetiology, mechanisms and epidemiology of chronic kidney disease (CKD) could help nephrologists better address the needs of their patients. Population-based studies indicate that CKD epidemiology differs by sex, affecting more women than men, especially with regard to stage G3 CKD. The effects of longer life expectancy on the natural decline of glomerular filtration rate (GFR) with age, as well as potential overdiagnosis of CKD through the inappropriate use of GFR equations, might be in part responsible for the greater prevalence of CKD in women. Somewhat paradoxically, there seems to be a preponderance of men among patients starting renal replacement therapy (RRT); the protective effects of oestrogens in women and/or the damaging effects of testosterone, together with unhealthier lifestyles, might cause kidney function to decline faster in men than in women. Additionally, elderly women seem to be more inclined to choose conservative care instead of RRT. Dissimilarities between the sexes are also apparent in the outcomes of CKD. In patients with predialysis CKD, mortality is higher in men than women; however, this difference disappears for patients on RRT. Although access to living donor kidneys among men and women seems equal, women have reduced access to deceased donor transplantation. Lastly, health-related quality of life while on RRT is poorer in women than men, and women report a higher burden of symptoms. These findings provide insights into differences in the underlying pathophysiology of disease as well as societal factors that can be addressed to reduce disparities in access to care and outcomes for patients with CKD.
The proportion of women with predialysis chronic kidney disease (CKD) is higher than that of men; this difference is likely due to the longer life expectancy of women and possibly to CKD overdiagnosis with use of estimated glomerular filtration rate equations
Kidney function declines faster in men than women, possibly owing to unhealthier lifestyles in men and the protective effects of oestrogens or the damaging effects of testosterone
More men than women start renal replacement therapy (RRT) not only owing to faster CKD progression in men but also because elderly women are more likely to choose conservative care
Mortality is higher among men at all levels of predialysis CKD, whereas mortality among individuals on RRT is similar for men and women
Women have reduced access to deceased donor transplantation compared with men, likely owing to higher levels of preformed antibodies, whereas access to living donor kidney transplantation in some countries seems equal
The perceived health-related quality of life of women on RRT is poorer than that of men, and women report a higher symptom burden and greater symptom severity than men
Many medical disciplines have become increasingly aware that diseases manifest differently in men and women. In cardiology, this phenomenon has been known for quite some time and has resulted in a number of medical innovations, such as the adaptation of coronary angiography diagnostics to the different pathophysiology of myocardial ischaemia in men and women, and alterations in emergency prioritization listing based on sex-specific symptoms of myocardial infarction1. However, differences in disease manifestations between men and women have not been so well explored in nephrology2,3. Differences in disease epidemiology, manifestation and outcomes can arise owing to biological (or sex) differences. However, differences can also arise owing to the sociocultural attributes of masculinity and femininity — known as gender differences — whereby men and women might be treated in a different manner or they might cope with their disease in a different way, perhaps by being construed to cultural and social behavioural expectations. Together, behavioural and biological differences between men and women can lead to differences in disease prevalence, progression rates and treatment outcomes, and this realization may de facto serve to identify new disease mechanisms and/or new and improved therapeutic opportunities. This Review discusses gender and sex differences in the epidemiology, treatment and outcomes of chronic kidney disease (CKD). We describe underlying reasons for these discrepancies where possible and also discuss possible explanations for some of the intriguing but unexplained disparities in CKD epidemiology between men and women.
Gender differences in CKD epidemiology
Prevalence of CKD. Estimations of CKD prevalence are central to the design of strategies to prevent and manage CKD at the population level. Data on the prevalence of CKD have grown dramatically over the past decade, building on the foundation set by standardizing the definition and staging of CKD in 2002 (Ref. 4) and contributing to our understanding of the scale of this public health problem. Despite these advances, however, the study of gender differences in CKD burden has received little attention. An analysis of population-based studies shows wide variation in the prevalence of CKD stages G3–G5 between countries5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27 (Fig. 1) (P. Trocchi, personal communication). The differences in CKD prevalence between countries might represent actual differences in CKD prevalence but might also be attributable in part to heterogeneity in data collection periods, variability in the use of equations to estimate glomerular filtration rate (GFR), issues with creatinine assay calibration, non-GFR determinants (such as differences in meat intake or muscle mass) and biases in population selection28. Despite this variation, in most geographical regions (with the exception of Japan and Singapore), the prevalence of CKD is higher among women than among men. The difference in CKD prevalence between men and women is also not constant between countries and is most evident in countries including France, Thailand, Portugal and Turkey, where the prevalence of CKD among women is twofold higher than in men.
The simplest explanation for the higher prevalence of CKD among women is that the longer life expectancy combined with the natural decline of kidney function with ageing contributes to an enlarged population at risk of CKD29,30. However, it is also possible that the use of equations to estimate GFR results in CKD overdiagnosis in women because these equations, such as the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) or the Modification of Diet in Renal Disease (MDRD) study equations, were not developed in community-based populations with appropriate age and ethnic diversity. A 2016 ancillary analysis of the Multi-Ethnic Study of Atherosclerosis (MESA)31 measured GFR (mGFR) by use of plasma clearance of iohexol in 294 unselected ethnically diverse and otherwise healthy adults (48% women). In agreement with earlier studies, they observed that women had lower mGFR than men (mean difference 21.39 ml/min; 95% CI 26.75–16.03 ml/min). However, this difference was largely attenuated after correction for body surface area (BSA) (mean difference after correction 9.34 ml/min; 95% CI 13.53–5.15 ml/min). In humans, BSA is the main predictor of kidney size. Men tend to have larger kidneys than women, and correcting estimated GFR (eGFR) by a constant BSA as we do currently with eGFR equations in clinical practice might therefore not be the best approach to achieve reliable values32. In a subsequent study33, the same researchers compared the performance of mGFR employing plasma clearance of iohexol with that of eGFR using the CKD-EPI equations (based on serum creatinine (eGFRcrea), cystatin C or the combination of creatinine plus cystatin C). Despite sex being included as a variable in the CKD-EPI formulae, the researchers found that eGFR underestimated mGFR in women more often than in men (for example, mean difference of mGFR-eGFRcrea 14.2 ml/min; 95% CI 16.5–10.9 ml/min in women versus 3.4 ml/min; 95% CI 6.3–0.0 ml/min in men). This larger measurement bias would lead to a consistent underestimation of eGFR in women in community-based studies and could explain the higher female CKD prevalence observed in various studies. Although this issue clearly deserves further study, it underscores a need for sex-specific thresholds in CKD classification systems, as has been discussed previously34,35.
Progression of CKD. Two large meta-analyses that reached opposite conclusions are often cited in discussions about the influence of sex on CKD progression36,37. In one, Neugarten et al.36 assessed the risk of CKD progression in 68 cohort studies of patients with nondiabetic CKD and concluded that men progress to end-stage renal disease (ESRD) faster than women. Subsequently, Jafar et al.37 performed a patient-level meta-analysis of 11 randomized trials that used angiotensin-converting enzyme inhibitors (ACEIs) in patients with CKD and concluded that the rate of renal disease progression might not be slower among women and in fact might even be faster among women than in men. The lack of agreement between these two meta-analyses can be attributed to several factors, including the mixed nature of the studies included (that is, the use of observational cohorts versus randomized controlled trials and population-based studies versus CKD referrals), the nature of patients included (for example, most women included in the later meta-analysis were postmenopausal, whereas the earlier study had a higher representation of premenopausal women) and the outcomes assessed (that is, slope of eGFR decline versus onset of renal replacement therapy (RRT)). Of note, the factors that influence these end points are different. Kidney function decline, for example, is largely driven by environmental, lifestyle and/or biological factors, whereas initiation of RRT is often influenced by nonbiological factors, such as access to health care and adequate provision of antiproteinuric medication. Most data from population-based studies (which reflect the population as a whole)38,39,40 suggest that men experience faster function decline (that is, a steeper slope of eGFR decline) than women. In the Dutch PREVEND study, for instance, men had a mean eGFR slope of −0.55 ± 1.47 ml/min/1.73 m2 per year compared with −0.33 ± 1.41 ml/min/1.73 m2 per year in women40. Various arguments have been put forward to explain this sex difference in GFR decline (Table 1), for example, protective effects of endogenous oestrogens versus deleterious effects of testosterone on kidney function and structure and the generally healthier lifestyles of women compared with that of men; however, clinical evidence to support these theories is rather weak.
Lifetime risk and incidence of renal replacement therapy. Several studies have indicated that the lifetime risk of RRT is higher among men than women. For a 40-year-old male in the USA, Canada and Europe, the lifetime risk of starting RRT for ESRD is 3.3% (in white individuals), 2.7% and 1.4%, respectively, whereas the lifetime risk of RRT for a 40-year-old female is 2.2% (in white individuals), 1.8% and 0.7%41,42,43, respectively. Thus, despite the higher prevalence of CKD in women, the majority of persons initiating RRT are men. This gender difference is surprisingly consistent across countries and time, with around 60% of those starting RRT for ESRD being men44,45.
The higher use of RRT among men supports the notion that progression of CKD occurs faster in men than in women and also reflects the fact that more women than men choose conservative care over RRT (discussed below). The faster progression of CKD in men would be expected to lead to a higher proportion of men with advanced stages of CKD than with earlier stages of CKD at a community level. Some support for this notion is provided by the Italian National Health Examination Survey46 and the Stockholm Creatinine Measurements (SCREAM) cohort14, in which 0.13% and 0.26%, respectively, of the male general population and 0.11% and 0.13%, respectively, of the female general population were found to have stage G5 CKD. However, some caution is needed in interpreting these findings because the numbers of patients with stage G5 CKD are relatively low, and the observed gender difference could therefore be due to random variation. More robust evidence for male dominance among the ESRD population is provided by an Australian study that estimated the total incidence of kidney failure (ESRD) between 2003 and 2007, including individuals who were not treated by dialysis or transplantation47. An incident case was defined as a newly registered case of treated ESRD on the Australian and New Zealand Dialysis and Transplantation Registry (ANZDATA) or a death registered with kidney failure recorded as a cause that was not recorded in ANZDATA (untreated kidney failure). The investigators showed that men comprised approximately 60% of the ESRD population and dominated across all age groups. Together, these data indicate that more men than women reach ESRD. Understanding what happens to the larger proportion of women in the community with milder forms of CKD would therefore seem to be of utmost importance. Women might simply have a slower progression of disease and therefore seldom develop ESRD (Table 1). However, one may also hypothesize that women are more likely than men to die before they reach ESRD or are not given the same opportunities to initiate RRT. These possibilities are discussed in further detail below.
In a Norwegian study that followed 3,047 patients with stage G3 CKD, the competing risks of death versus development of ESRD (defined as an eGFR <15 ml/min/1.73 m2 or initiation of RRT) were modelled over 10 years of follow-up38. The researchers observed higher risks of mortality and ESRD for men than for women (10-year cumulative incidence of death 0.61 and 0.47 and of kidney failure 0.08 and 0.03 in men and women, respectively). This higher risk of death among men was confirmed by a meta-analysis of >2 million individuals from multiple international cohorts; this analysis by the CKD Prognosis Consortium demonstrated that mortality was overall higher in men than in women at all levels of renal function (as assessed by eGFR and albumin:creatinine ratio)48 (Fig. 2). It also showed that the slope of the risk relationship between all-cause mortality and eGFR decline was steeper in women than in men, indicating that the increase in mortality risk associated with CKD progression was greater for women than for men. Furthermore, the association between eGFR and mortality risk differed depending on the type of cohorts studied. Although the mortality of men was higher than that of women among population-based cohorts across all eGFR thresholds, this sex difference in mortality was abolished among cohorts of CKD-referred patients; the researchers attributed this lack of difference in CKD-referred cohorts to selection bias (that is, women with CKD dying before referral or being referred later than men) or to renal care being more equal for women and men than within primary care48. Thus, although available evidence points to a sex difference in mortality risk associated with eGFR in the community, the risk of death associated with further reductions in eGFR might not differ by sex following referral to a CKD clinic.
As mentioned above, differences in RRT initiation between men and women might not only relate to differences in the rate of CKD progression but could also relate to nonbiological factors, such as access to care and personal preference. Several studies have demonstrated that the proportion of patients with ESRD on conservative care (that is, patients not initiating RRT) increases sharply at >60 years of age47,49 and that among patients 70–75 years of age, women are less likely to receive RRT than are men47,50. This finding is consistent with previous reports showing that elderly women, when referred to nephrologists, are two to three times more likely than elderly men to choose conservative care instead of RRT51,52. Reasons for this difference could relate to the fact that more elderly women than men live alone and lack a caregiver — a role often fulfilled by spouses47. Other factors could relate to gender differences in access to nephrology care. For instance, a Swedish study found that although women are more likely to have CKD, they are less likely to have received an International Classification of Disease, Tenth Edition (ICD-10) CKD diagnosis or to have consulted a nephrologist14. In the USA, women in community populations are reportedly less aware of their CKD than men53, and women starting dialysis are more likely to belong to socioeconomically deprived minority groups and are more likely to be uninsured or unemployed than men54,55. If and to what extent sociocultural and health-care factors might affect sex differences in the lifetime risk of ESRD and initiation of RRT remain unknown.
Dialysis initiation. Studies have shown that women start dialysis at eGFR levels that are on average slightly lower (0.6–0.8 ml/min/1.73 m2) than those of men45,56,57. These differences might reflect actual differences in the timing of dialysis initiation but, of note, might not be clinically relevant and might not be particularly accurate given the known caveats of eGFR equations. Women who start dialysis are on average 1–2 years older than men44,45. Interestingly, the proportion of men and women receiving pre-ESRD nephrology care in the USA before RRT initiation is remarkably similar, according to the latest report from the US Renal Data System (USRDS) (34.7% and 35.3%, respectively), with similar duration of nephrology care45. In terms of dialysis modality, no apparent gender differences have been reported. In the USA, 88% of men and women started RRT on haemodialysis, whereas peritoneal dialysis was the modality of choice in 9% of patients regardless of sex45; in Europe, 82% of men and 81% of women started RRT on haemodialysis, whereas 14% of men and women started on peritoneal dialysis44.
At initiation of RRT, men have more comorbidities than women, especially cardiovascular disease and cancer45,58,59, and are more likely to be smokers60, both of which are indicative of unhealthier lifestyles and are in line with the lower healthy life expectancy of men and the higher mortality of men on RRT in some cohorts. Sex-specific complications are also evident, explained in part by sex differences in age of CKD onset, by the differences in the prevalence of common risk factors for CKD and by the direct consequences of kidney failure, for example, on luteinizing hormone signalling and prolactin retention, with subsequent inhibition of gonadotropin secretion61. Testosterone deficiency is present in 40–50% of men undergoing RRT62 and probably contributes to an overall state of catabolism, being associated with anaemia, protein–energy wasting, osteoporosis, cardiovascular complications and renal transplant rejection63,64,65,66. Similar issues arise in women undergoing RRT, in whom oestrogen deficiency is also very common, accelerating menopause by 4–5 years on average, and is linked to sleep disorders, depression, osteoporosis, impaired cognitive function and increased cardiovascular risk67.
Quality of dialysis and CKD care. Although available data, at least from the USRDS, suggest that no gender difference exists in access to or duration of predialysis nephrology care44, a number of aspects of dialysis care might differ. For instance, women are more likely than men to receive dialysis for <12 h per week68. Estimations of dialysis adequacy are based on Kt/V, which assumes that V, urea distribution volume, is constant for all individuals69. Because urea distribution volume is a surrogate for lean body mass, such an assumption obviates that women, in general, have lower muscle mass than men. It has therefore been suggested that Kt/V overestimates dialysis adequacy in women70, and the unexpected observations from the HEMO study71,72,73 showing that a high dialysis dose leads to lower risk of mortality among women but not among men have been attributed to such overestimations74.
Arteriovenous fistulas (AVF) are considered the first choice of vascular access for patients on haemodialysis. However, the majority of patients start haemodialysis using a catheter. In both Europe and the USA, the use of catheters at initiation of dialysis is slightly more common in women than in men (62.9% versus 59.3% in Europe and 81.5% versus 79.4% in the USA)45,75. Concerns about the smaller vascular diameters of women might prompt nephrologists to consider use of a catheter over an AVF; however, duplex ultrasonography studies have demonstrated that the vasculature of female patients is as adequate as that of men for the placement of AVFs76. Some researchers suggest that the small gender difference in catheter use is therefore partially explained by women opting more often for catheters instead of AVFs owing to cosmetic reasons77. Catheter use in prevalent dialysis patients is lower than that in incident dialysis patients but remains higher in women than men (32.3% versus 25.6% in Europe and 21.2% versus 16.9% in the USA)45,75. The greater gender difference in catheter use among prevalent patients compared with incident patients might be due to the higher percentage of failed AVF placements, their lower maturation rate and longer time to first use among women45.
Beyond differences in dialysis dose and access, differences might also exist in the pharmacological treatment of dialysis-dependent patients with CKD. For example, definitions of anaemia in patients with CKD might contribute to overestimations and overtreatment of anaemia in women78,79,80. In the general population, women have lower haemoglobin levels than men, and anaemia in the general population is defined by sex-specific thresholds81. Sex differences in haemoglobin levels are not, however, considered in CKD guidelines, which recommend use of the same haemoglobin target for men and women82,83. In line with this observation, women on dialysis require higher doses of erythropoietin-stimulating agents than men in order to achieve the same haematocrit concentration84,85,86.
Access to transplantation
Access to a deceased donor organ. The realization that access to kidney transplantation is lower for women than for men dates back to 1988, with the publication of several prominent analyses87,88,89. Deceased donor transplantation involves a stepwise process90, whereby patients (who are usually on dialysis) undergo a series of evaluations before being placed on the wait list for a deceased donor organ. Whether a transplantation takes place then depends on the availability of a well-matched organ in terms of blood group, HLA antigens and preformed antibodies.
As fewer women than men start dialysis, the finding that a smaller absolute number of women than men receive deceased donor kidneys does not come as a surprise. Analyses from the 1990s, however, showed that not only absolute numbers of transplants but also transplantation rates were lower among women than among men91,92 and that these differences could not be explained by comorbidities and other patient characteristics. A 1997 study from the USA showed that women were both less likely to be on the transplant waiting list and, once on the list, less likely to receive a deceased donor renal transplant than men93. Similar sex inequalities in kidney transplantation rates were also reported in a 2000 study from Canada94, whereas a large US analysis from 2009 suggested that there was no disparity in access to transplantation for women in general but rather a marked disparity in access to transplantation for older women and women with comorbidities, despite similar survival benefits from transplantation for men and women regardless of age or comorbidities95.
A seminal paper from 2000 (Ref. 96) analysed sex-dependent differences throughout the process of wait-listing and transplantation by use of data from the USRDS96. The researchers found that among patients who started ESRD treatment from 1991–1996, fewer women than men were placed on the transplant wait list, which was equivalent to a 16% lower rate for wait-listing. Although the reasons for this difference are unclear, it might in part be due to patient preference, gender selection bias by health-care personnel or on the part of family and friends, or socioeconomic reasons96 (Fig. 3). The finding that women received 14% fewer transplants than men after wait-listing was, however, fully explained by higher levels of preformed lymphocytotoxic antibodies among women transplant candidates, as transplantation rates were not significantly different between men and women after adjusting for preformed lymphocytotoxic antibody level96. Although the most recent USRDS report has shown that the gap in unadjusted deceased donor kidney transplantation rates between men and women has narrowed since 1997, an analysis of deceased donor kidney transplantation rates between men and women should be expanded with data from outside the USA.
Access to living donor transplantation. Many studies have reported that women are more often living kidney donors than recipients of a living donated kidney97,98,99,100,101,102 — a finding that would ad hoc be considered unfair. However, this issue is not easily understood and may not be as unfair as it seems. Again, a key consideration relates to the fact that fewer women than men initiate dialysis, and if judged solely by ESRD incidence, fewer women than men might be expected to need a renal transplant. According to the Organ Procurement and Transplantation Network (OPTN) and Scientific Registry of Transplant Recipients (SRTR) annual report, women comprised 40.7% of wait-listed kidney transplant candidates in the USA in 2012 and received 38.8% of all kidney transplants, including 39.3% of all deceased donor transplants and 37.5% of all living donor transplants103. These data suggest that the rate of transplantation for women is proportional to the percentage of women who are wait-listed. Of note, the percentage of living donor kidney transplantations relative to the total number of kidney transplantations for women is very similar to that for men over the past decade (a mean of 37% annually for both sexes), suggesting that there is not a sex disparity in living donor kidney transplantation rates despite notable barriers104.
If one assumes that females are simply healthier than males, then the finding that more women than men are living organ donors should not be striking. The situation, however, is of course more complex. An assessment98 of the actual and expected sex distribution of living donors and recipients of living donor kidneys aged <65 years of age from US census data found that the differences between the observed and expected proportions were skewed towards fewer living unrelated non-spousal donations for women (observed versus expected Χ2 = 382.7 for living related donation and Χ2 = 37.0 for living unrelated non-spousal donation, both P < 0.0001 (Ref. 97)). The researchers suggested possible reasons for this disadvantage, including a higher degree of ambivalence about organ donation in men105, a greater incidence of coronary artery disease and hypertension among men (eliminating a greater proportion of males from the potential donor pool)91, potential unavailability of the men as a result of military obligations or incarceration and financial disincentives, such as the absence of a guaranteed reimbursement system for lost wages98; however, the causality of these associations remains inconclusive.
As with deceased donor transplantation, data and high-quality analyses for living kidney donor transplantation from outside the USA are scarce. Previous review papers106,107,108, as well as our search of reports from China109, Egypt110, India111,112, Iran113,114,115,116, Korea117, Nepal118, Nigeria119, Saudi Arabia120 and Tunisia121, demonstrate that worldwide, women donate more frequently than men, with the exception of Iran, where organ donation is driven by economic factors106,115. Importantly, however, the percentage of female recipients of a living donor transplant across these studies was lower than the 37.5% reported for the USA103. Specifically, the percentages of female recipients reported in these mostly smaller studies were 20.5% for China109, 26% for Egypt110, 11%111 and 14%112 for India, 36.9%113, 29.8%114 and 38.14%116 for Iran, 31% for Korea117, 29% for Nepal118, 23% for Nigeria119, 35% for Saudi Arabia120 and 34% for Tunisia121. By contrast, the situation in England (single centre experience122), Germany (national registry data123), Russia (single centre experience124) and Switzerland (national registry data125) seems more similar to that of the USA (with living donor transplantation rates among women at 39.4%122, 37.6%123, 42.4%124 and 36%125, respectively). Interestingly, data from the national registry of Thailand demonstrate this country to be an exception to the trends described above: although organ donation in Thailand is not paid for, 50.7% of living donor kidneys were donated by men from 1987–2012, whereas the rate of transplantation into female recipients was comparable to that of the U.S. (37.7%)126. The reasons for these international differences remain altogether speculative and demand further study, preferably from national registries rather than single centres.
Outcomes of renal replacement therapy
Sex differences in quality of life
Similar to findings in the general population127,128 and individuals with atrial fibrillation or HIV/AIDS129,130, perceived health-related quality of life (HRQOL) is poorer among women on RRT than in men on RRT across both mental and physical domains131,132,133,134,135,136,137 (Fig. 4). Women on dialysis report a higher symptom burden and greater symptom severity than do men135,138,139,140,141,142 and take longer to recover after a dialysis session than men141. Transplantation improves perceived HRQOL in both sexes, but this improvement is less pronounced in women143,144. Moreover, the negative effects of graft loss on HRQOL are greater in women than in men145.
The gender disparity in HRQOL has been attributed to the different ways in which men and women experience and react to ESRD. First, depression is more often diagnosed in women on RRT than in men58,134,146, and depression prevalence correlates directly with uraemic symptom burden and severity147. The higher prevalence of depression in women has been suggested to explain a substantial part of the gender differences across various HRQOL domains142. Of interest, the effect of depression in explaining gender differences in HRQOL has been noted in cardiology129,148 and in adolescents with diabetes mellitus149,150. Second, psychosocial differences in how men and women adapt to RRT might affect their perceived HRQOL. Women might make more use of emotional and social-support-seeking strategies to cope with their disease, whereas men tend to adopt a more problem-solving-oriented approach151. Men have also been noted to more often adopt avoidance mechanisms (for example, by heavy drinking and smoking) as a coping strategy, although how this approach influences HRQOL is unclear151. Women report higher stress in response to physical symptoms of disease (such as nausea, vomiting, muscle cramps, joint stiffing, fatigue and loss of bodily function), whereas men perceive themselves as better able to cope with the physical aspects of their disease151,152,153,154. Third, men and women might receive different levels of social support. This observation has previously been made in populations with cardiovascular disease155,156 but remains speculative in patients on RRT132,137. Men on RRT might receive more social support than women owing to socially determined gender roles (for example, men on RRT more often are married than women on RRT58 and might be taken care of by their wife and family more often than women are taken care of by their family) and to the lower life expectancy of men. Finally, owing to differences in socially and culturally determined gender roles, men and women might differ in their perception of illness, attribute different values to various aspects of health133, differ in their reporting, description, and labelling of disease and differ in their inclination to disclose discomfort127.
Sex differences in hospitalizations
Although hospitalization is an inherently limited outcome measure in aetiological and prognostic research, it is fundamental for health-economy research and for research into the quality of clinical care. In the USA, women on haemodialysis have an approximately 20% higher rate of all-cause hospital admissions and all-cause hospital duration than men157,158. Similarly, women on peritoneal and haemodialysis have a higher risk of infection-related hospitalization than men (24% and 18%, respectively)159,160. Furthermore, among US adults placed on the waiting list for a first deceased donor kidney transplant, women have an 11% higher risk of hospitalization than men161. A 2017 study reported that the difference in hospitalization rate between women and men on haemodialysis decreases with increasing age. Among patients aged 18–34 years, women had a 54% greater risk of hospitalization than men, whereas this risk was reduced to 16% among patients >75 years of age. In that study, the higher hospitalization rate in women was mostly explained by lower serum albumin levels, which likely reflected their poorer health status158. Other factors that have been suggested to explain this gender discrepancy include medication and/or therapy nonadherence, greater severity of illness and a lower likelihood to receive care as recommended by guidelines in women162.
Sex differences in death and causes of death
The expected lifespan of a dialysis patient <80 years of age is less than a third of that of an age-matched healthy individual in the general population. The life expectancy of transplant recipients is somewhat better (45–85%) than that of an age-matched healthy individual44,45. In both dialysis patients and transplant recipients, however, the survival advantage of women in the general population (around 4 years at the age of 50 and 3 years at the age of 70) is reduced to only a few months for patients on dialysis and up to a year for transplant recipients44. These data are in line with the progressive loss of the female survival advantage observed with decreasing renal function in patients with CKD48.
As a result, the overall patient survival of men and women on RRT is very similar163 (Fig. 5). Nevertheless, sex differences emerge within different age categories, with diabetes status and with cause of death (Tables 2,3). Below 45 years of age, women starting dialysis have a higher mortality risk than men, but at older ages, they have a survival advantage. The higher risk of death among younger women is attributable to noncardiovascular causes; by contrast, in women with diabetes mellitus, the increased risk of death from noncardiovascular causes is elevated compared with that of men across all age categories163. Various studies have shown that the higher noncardiovascular death among younger women and those with diabetes mellitus is predominantly attributable to infection-related mortality77,164. One study of patients on peritoneal dialysis reported that the risk of death from infection in women is almost double that of men and is related to a higher occurrence of death due to sepsis and peritonitis164. An ERA-EDTA Registry study also demonstrated that young female patients receiving dialysis and transplant recipients are prone to death from infections77. This finding can be partially explained by a higher prevalence of multisystem disease in young female patients, particularly systemic lupus erythematosus or the earlier described higher use of catheters for vascular access in women77. Together, these data suggest that young women with diabetes mellitus and multisystem disease deserve particular attention in terms of preventing infections.
Other sex differences in cause of death are evident among older patients on dialysis. Men >45 years of age initiating dialysis — particularly those without diabetes mellitus — are at higher risk of cardiovascular mortality than similarly aged women on dialysis163, which is in line with the higher prevalence of cardiovascular comorbidities among this population. In addition, male dialysis patients and transplant recipients aged >40 years die more frequently from cancer than do their female counterparts77. This finding is concordant with data from the general population where women have a survival benefit for a number of tumours, such as colorectal cancer and melanoma165,166. Finally, findings from registry studies indicate that women have an 18% higher risk of withdrawal from dialysis than do men167,168. Although gender disparities in clinical decision-making and societal value judgements have been suggested to cause this difference168, further investigation is warranted.
Kidney allograft survival
The probability of all-cause kidney allograft failure is lower for living donor transplants than for deceased donor transplants and, according to US data, continuously improved between 1991 and 2012 (Ref. 103). Several fundamental questions need to be addressed in terms of the effects of gender on kidney allograft survival.
Sensitization during pregnancy. First, what is the effect of sensitization during pregnancy on allograft outcomes? This question has been addressed in a study169 that analysed the graft survival rates of kidneys donated by spouses, living unrelated donors, parents, HLA-identical siblings, offspring and cadavers. Kidney graft survival was best for HLA-identical siblings, followed by parental donor grafts with one HLA-haplotype mismatch; cadaveric donation was associated with the worst allograft outcomes. The fact that kidney grafts from living unrelated donors had higher survival rates than cadaveric grafts despite a higher degree of HLA mismatching was attributed to potential damage of the cadaveric kidneys (before removal) rather than the length of cold ischaemia time169. Of note, however, the researchers found that women who had previously been pregnant and received a living kidney from their spouse were at greater risk of long-term graft failure than were men who received a graft from their spouse or women who received a graft from their husband but had not previously been pregnant (3-year graft survival rate of 76% for women with previous pregnancies compared with 87% for men and for women with no previous pregnancies). This finding supports previous studies showing that HLA sensitization frequently occurs in multiparous women170,171 and is in line with a description of humoral rejection in a multiparous recipient of a graft donated by her living spouse172. The finding of humoral rejection — characterized at that time by peritubular capillaritis and linear C4d deposition within peritubular capillaries173 — in the rejected graft of a multiparous recipient172 identified a link between humoral rejection174 and HLA antibody formation during pregnancy175,176. The clinical consequences of such HLA formation are highlighted by a 2017 single centre study that compared the ability of men and women to navigate the different steps of living donor kidney transplant evaluation and transplantation (that is, referral, donor-specific histocompatibility testing at actual transplantation); the researchers found that female candidates fall behind their male counterparts during histocompatibility testing, leading to reduced transplantation rates despite similar rates of referral from the wait list (which contained 38.4% women)177.
Sensitization during development and following birth. A related question is the role of the sensitization process following exposure to maternal antigens during development, upon delivery and with nursing. The sensitization process that occurs in a child following exposure to maternal antigens in utero and upon nursing seems to induce tolerance to foreign antigens178. This notion is supported by the findings of a notable study of transplant recipients who received kidneys from sibling donors who were mismatched for one HLA haplotype. In that study, graft survival was higher (although early rejection episodes were also higher) when the donor had maternal antigens not inherited by the recipient instead of paternal antigens178. Although these findings support the hypothesis that tolerance to a later antigen challenge is induced by cells and antigens of the mother179,180, an analysis of data from the ANZDATA registry reported contradictory findings181, whereby maternal donor kidneys were associated with poorer graft outcomes than paternal donor kidneys owing to an increased number of rejection episodes.
Sex differences in kidney allograft function and outcome. Despite greater understanding of immunological processes that occur in transplantation, we still do not completely understand the degree to which sex differences influence kidney allograft outcomes and allograft function. The most recent thorough analysis of the topic182 shows that the risk of kidney allograft failure is generally higher or equal in women than in men, with the exception of women ≥45 years of age, who have a lower risk of graft failure than their male counterparts. The sex of the donor organ can lower the risk of allograft failure in women, with female kidneys being more favourable. This study182 as well as others have revealed a complex interplay of factors that might contribute to these sex differences in graft outcomes, including differences in immune reactivity to sexually determined minor histocompatibility antigens183,184,185, age-related differences in the immune reactivity to foreign antigens186, the effect of sex hormones on immune activation processes187,188, differences in metabolic demands due to sex-related differences in body size189,190 and sex differences in adherence to immunosuppressive treatment191,192, although not all studies have shown that women are always more adherent than men193. These studies demonstrate that understanding the effects of sex on kidney allograft survival can be achieved only by also considering the effects of age. They also highlight that further study is needed to elucidate the underlying mechanisms.
The data described here demonstrate a number of differences in the epidemiology and outcomes of men and women with dialysis-dependent and nondialysis CKD. Whereas differences in the progression of CKD and initation of RRT might have a biological basis, differences in the choice of treatment for ESRD might be explained by sociocultural factors. For instance, as we have described, elderly women seem to be more inclined than elderly men to choose conservative care instead of RRT. This finding, together with apparent gender differences in wait-listing to the disadvantage of women, could potentially lead to reduced access to care for women. Moreover, why women have greater severity of uraemic symptoms than men and a lower perceived quality of life is not well understood. Although the knowledge base on gender differences in the CKD population has grown over the past couple of years, we strongly encourage future efforts to continue exploring the existence and underlying mechanisms of gender differences in CKD epidemiology, access to treatment and outcomes. Through this approach, we will be able to identify sex-specific and gender-specific adaptations and innovations in medical practice that might lead to personalized and improved patient care.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Nature Reviews Nephrology thanks P. Delanaye, P. Eggers and E. Ku for their contribution to the peer review of this work.
The authors thank P. Trocchi (Universitätsklinikum Essen, Germany) for providing sex-specific statistics from Germany and F. K. Port (Arbor Research Collaborative for Health, Ann Arbor, Michigan, USA), as well as T. Stamm, G. Böhmig and G. Bond (all from Medical University of Vienna, Austria) for their helpful comments and revisions to this work. J.J.C. acknowledges grant support from the Swedish Heart and Lung Foundation and the Westman and Rind foundations. N.C.C. and K.J.J. acknowledge grant support from the European Renal Association–European Dialysis and Transplant Association (ERA-EDTA).
The preferred method for measuring the dialysis dose; defined as the dialyser clearance of urea (K) multiplied by the duration of the dialysis treatment (t, in minutes) divided by the volume of distribution of urea in the body (V, in ml), which is approximately equal to total body water, corrected for volume lost during ultrafiltration.
- Prevalent dialysis patients
All patients treated by dialysis at a particular moment in time.
- Incident dialysis patients
Patients starting dialysis for the first time.
- HLA sensitization
Formation of alloantibodies against HLA antigens.