Chapter Outline
EPIDEMIOLOGY OF END-STAGE KIDNEY DISEASE, 638
EPIDEMIOLOGY OF CHRONIC KIDNEY DISEASE, 643
Prevalence, 645
Incidence, 645
Outcome by Stages, 645
International Comparisons, 647
Relationship Between CKD Epidemiology and ESKD Epidemiology, 647
EPIDEMIOLOGY OF ACUTE KIDNEY INJURY, 647
LINKING ACUTE, CHRONIC, AND END-STAGE KIDNEY DISEASE, 652
Epidemiology is defined by the Oxford English Dictionary as “the study of the incidence and distribution of diseases.” This chapter focuses on the epidemiology of end-stage kidney disease (ESKD), chronic kidney disease (CKD), and acute kidney injury (AKI), the three most important clinical problems in nephrology, measured either by number of patients affected or by rates of associated morbidity and mortality. Some initial definitions are useful. The incidence rate of a disease is typically defined as number of new cases per person-year from longitudinal studies that enrolled patients without disease at baseline. The prevalence rate of a disease is typically defined as number of persons with disease per population at any one point in time in cross-sectional studies. Disease prevalence thus depends not only on disease incidence but also how long the condition persists. Both incidence rates and prevalence rates are typically normalized to some underlying population (e.g., the total U.S. population). Hence, if the underlying population size were to increase, as the U.S. population has, absolute incidence (and prevalence) count could increase over time even if the incidence (or prevalence) rate were unchanged.
This chapter also touches upon risk factors for kidney disease, which are discussed in greater detail in Chapters 21 and 22 .
Epidemiology of End-Stage Kidney Disease
In the study of kidney disease epidemiology, the traditional focus has been on end-stage kidney disease (ESKD), usually defined operationally as kidney failure being treated with long-term dialysis or kidney transplantation. ESKD is the most serious and dramatic manifestation of kidney disease and naturally was the focus of much clinical attention in the early decades of nephrology, the 1950s and 1960s. During this time, with pioneering work in renal transplantation and dialysis, nephrology became recognized as a separate discipline. Several decades’ worth of very strong data exist with regard to the epidemiology of ESKD, in large part because of the existence of ESKD registries such as the Michigan Kidney Registry and United States Renal Data System (USRDS). USRDS provides powerful epidemiology data because it has nationally comprehensive patient-level data and tracks outcomes after the diagnosis of ESKD.
Incidence
The USRDS Annual Data Report for 2013 showed for the first time that not only had the incidence rate of ESKD fallen (by 3.8% to 357 cases per million people in 2011, the most recent year for which data were available), the incident count also fell from 117,390 (in 2010) to 115,643 (in 2011) ( Figure 20.1 ). This was the first time since 1980 that USRDS ever documented a fall in the absolute number of new (incident) ESKD cases in the United States.
This secular trend is quite different from that observed in the 1980s and 1990s, when the incidence rate of ESKD rose rapidly ( Figure 20.2 ). It is, however, a continuation of the trend reported by USRDS that the incidence rate of ESKD has more or less reached a plateau in the past decade (see Figure 20.2 ).
Similar encouraging observations have been reported from other countries. For example, the 2014 Canadian Organ Replacement Register (CORR) Report, the 2013 UK Renal Registry’s Annual Report, and the 2013 Australia and New Zealand Dialysis and Transplant Registry (ANZDATA) Report all show stable or declining incidence of ESKD in their respective countries ( Figure 20.3 ).
The reasons for these encouraging trends are not entirely clear but may involve successes in retarding the progression of CKD resulting from more aggressive control of blood pressure with drugs such as those that block the renin angiotensin aldosterone system (RAAS). For example, when the landmark United Kingdom Prospective Diabetes Study (UKPDS) trial was conducted in the 1980s, it was thought acceptable to allow patients with type 2 diabetes in the control arm to have blood pressures as high as 200 mm Hg systolic/105 mm Hg diastolic (this value was lowered to 180/105 mm Hg in 1992). The first large, randomized controlled trial to demonstrate the renoprotective effect of RAAS blockade was not published until 1993. In addition, improved glycemic control among patients with diabetes mellitus may have an important role. The impact of such treatment is illustrated by a Finnish study that tracked outcome among young patients with type 1 diabetes over several decades. In that study, patients with type 1 diabetes diagnosed from 1980 through 1999 had less than half the risk for development of ESKD of those diagnosed from 1965 through 1969. Similar results have been reported in the United States for patients with type 2 diabetes, with one paper reporting a 28% drop in rates of ESKD between 1990 and 2010.
There remains considerable variation in the incidence of ESKD across patient subgroups. In the United States, the incidence of ESKD among African Americans is strikingly higher than the incidence among whites (by more than threefold after adjustments for gender and age) ( Table 20.1 ). Incidence rates are also higher among those of Asian descent and Hispanic ethnicity (see Table 20.1 ). Men (compared with women) and older (compared with younger) people have higher rates of ESKD (see Table 20.1 ).
| Subgroup | Adjusted Incidence (per million population) |
|---|---|
| By Race | |
| White | 279.8 |
| Asian | 398.5 |
| Native American | 452.5 |
| Black/African American | 939.8 |
| By Ethnicity | |
| Non-Hispanic | 342.7 |
| Hispanic | 517.5 |
| By Age (yr) | |
| 0-19 | 15.6 |
| 20-44 | 126.5 |
| 45-64 | 571.1 |
| 65-74 | 1306.8 |
| 75+ | 1706.9 |
| By Gender | |
| Female | 281.3 |
| Male | 451.2 |
* Overall incidence (adjusted for age, gender, and race) is 357.0 per million population. Rates by race or ethnicity are adjusted for age and gender; rates by age are adjusted for gender and race; rates by gender are adjusted for age and race.
The most common listed etiology of ESKD in the United States is diabetes (156.8 per million population; 44%), followed by hypertension (100.6 per million population; 28%).
However, these are “primary diagnoses” for ESKD assigned by the treating physician at the start of dialysis, when the original etiology of disease may be difficult to discern. For example, it has been documented that numerous cases of ESKD ascribed to hypertension have preceding clinical features highly suggestive of other renal parenchymal diseases (e.g., nephrotic range proteinuria). Also, genetic studies have demonstrated that variants in the gene encoding apolipoprotein L1 (APOL1) appear to account for much of the higher incidence of “hypertensive nephrosclerosis” observed among African Americans. These novel advances will undoubtedly force a rethinking of how we currently assign etiologies for ESKD—specifically the question, “Should we accept hypertension as being the cause of ESKD in a large number of cases?”
Furthermore, the convention of ascribing only one primary cause to any ESKD case may be inherently limited. For example, prospective studies have shown that patients with diabetes mellitus appear to be at several-fold higher risk for ESKD ascribed to nondiabetic causes. This convention also cannot reflect the contribution of multiple disease processes to the final ESKD outcome (e.g., nonrecovery of renal function after acute tubular necrosis in a patient with underlying diabetic nephropathy).
Prevalence
According to the 2013 USRDS Annual Report , the annual prevalence of ESKD in the United States was 1901 per million population (adjusted for age, gender, and race). As alluded to previously, the prevalence of disease depends not only on the number of new cases but also on the survival of existing patients. Better survival of patients on both dialysis and transplant will therefore increase the prevalence of disease (and “burden of ESKD”) but actually reflects improvement in care. Indeed, the adjusted mortality rates for patients undergoing dialysis and transplantation have improved over the past three decades ( Figure 20.4 ).
Another reflection of the difference between incidence and prevalence is that although only 3% of patients with incident ESKD are treated by kidney transplantation as the initial modality, 30% of patients with prevalent (existing) ESKD are maintained with kidney transplants. This is due not only to the fact that numerous patients undergoing dialysis subsequently undergo kidney transplantation but also because patients with kidney transplants survive longer. Similarly, the black-white disparity in ESKD is even more pronounced when ESKD prevalence is used as the metric rather than ESKD incidence, because on average, black patients survive longer with ESKD than their white counterparts.
Mortality rates among patients with ESKD, especially those undergoing dialysis, remain alarmingly high—in excess of 20% per year (see Figure 20.4 ). This high mortality rate is paralleled by high rates of hospitalization and health care utilization. Several studies have shown that death rates are particularly high in the first weeks to months immediately after patients start hemodialysis, and there is concern that this problem is underestimated by some national registry data because patients die prior to being registered.
Many epidemiology studies have sought to account for the high mortality rates among patients undergoing dialysis. The burden of medical conditions already present at the start of dialysis appears to be a key problem. This possibility is consistent with the observation that interventions to manipulate dialysis-related parameters, such as dose of dialysis and use of more recombinant erythropoietin, have not succeeded in reducing mortality.
Notably, many papers have reported that numerous risk factors for mortality in the general population, such as higher blood pressure, higher cholesterol level, and higher body mass index, appear to be paradoxically associated with lower risk of mortality among patients receiving maintenance hemodialysis. The reasons for these reverse “J-shaped” or “U-shaped” associations are not entirely clear. Part of the explanation may be confounding factors such as malnutrition and inflammation. Another factor may be the unique physiology of patients undergoing hemodialysis, such as hemodynamics related to interdialytic fluid accumulation, because several studies have shown that higher blood pressure measured outside the dialysis unit (in contrast to measured just before a hemodialysis session ) was associated with a linear increase in the risk of adverse outcomes.
Results of interventional studies are mixed, with some but not other studies showing benefits of treating conventional “Framingham” cardiovascular disease risk factors after onset of ESKD, such as lipid lowering. And there is some evidence (but not strong) in favor of blood pressure lowering in the dialysis population.
Kidney Transplantation
Although no randomized controlled clinical trials have been performed, the best evidence from observational data indicate that, all else being equal, receipt of a kidney transplant confers mortality benefit in addition to improved quality of life in comparison with maintenance dialysis. Currently, 1-year survival with a functioning allograft is 92% for recipients of deceased donor kidneys and 97% for recipients of living donor kidneys.
In 2011, the number of kidney transplantations performed in the United States was 17,671. Figure 20.5 shows the trend over time according to donor type. Shortage of transplant organs, however, remains a major problem that has resulted in longer waiting times. The pressure to increase number of organs has led to numerous contentious discussions regarding proposed schemes to pay donors and use of living donors who have medical conditions that were previously considered contraindications to donation, such as hypertension.
International Comparisons
Incidence and prevalence rates of ESKD vary considerably across different regions and countries. In 2011, Jalisco (Mexico), the United States, and Taiwan (2010 figure) reported the highest rates of incident ESKD at 527, 362, and 361 per million population, respectively ( Figure 20.6 ). Interpreting incidence rates of ESKD across countries however, is complicated by lack of uniform data collection methods. Furthermore, because only treated ESKD cases are counted, differences in access to renal replacement therapy due to economics, public policy, or local practice patterns greatly influence incidence and prevalence rates.
Despite these limitations, it is clear that there is considerable variation in the practice of renal replacement therapy around the world. For example, in Hong Kong, the great majority of patients undergoing dialysis are treated with peritoneal dialysis (74% of prevalent patients [those continuing dialysis] vs. 33% in New Zealand, 20% in Iceland, and 7% in the United States). Home hemodialysis use also varies greatly (18% in New Zealand, 9% in Australia, and 1% in the United States). In terms of transplantation, the reported rates in 2010 were 57 per million person-years in the United States, 37 in Israel, 29 in Argentina, and 8 in Romania.
Finally, the economic models of delivery of dialysis also vary. In the United States currently, about two thirds of patients with ESKD are dialyzed at facilities owned by large dialysis companies. In other countries (such as Germany), the system is much more decentralized.
Beyond registry data, international observational studies such as the Dialysis Outcomes and Practice Patterns Study (DOPPS) have documented noticeable practice variations. For example, in the late 1990s, arteriovenous fistula was used as the dialysis access in 80% of European but only 24% of U.S. prevalent patients. For patients who were new to hemodialysis, the rates of fistula use were 66% in Europe and 15% in the United States. This and other differences may explain the much higher mortality among U.S. patients than patients elsewhere in the world. For further international comparisons, see Chapter 77 , Chapter 78 , Chapter 79 , Chapter 80 , Chapter 81 , Chapter 82 .
Epidemiology of Chronic Kidney Disease
The past dozen years have seen an explosion of research into CKD. This field was codified in large part following the publication in 2002 of the National Kidney Foundation (NKF) Kidney Disease Outcomes Quality Initiative (KDOQI) definition and classification of chronic kidney disease. Prior to this, there had been no consensus definition of CKD, a lack that made defining disease incidence and prevalence difficult. The NKF CKD definition and classification have been very influential, widely disseminated, and widely adopted ( Table 20.2 ). Although a newer classification of CKD—building on the KDOQI definition but with additional emphasis on the role of albuminuria—has since been put forth by the NKF program Kidney Disease/Improving Global Outcomes (KDIGO), in 2011 it has not yet influenced studies of the population epidemiology of CKD.
| Stage | Description | GFR (mL/min/1.73 m 2 ) |
|---|---|---|
| 1 | Kidney damage with normal or ↑ GFR | ≥90 |
| 2 | Kidney damage with mild ↓ GFR | 60-89 |
| 3 | Moderate ↓ GFR | 30-59 |
| 4 | Severe ↓ GFR | 15-29 |
| 5 | Kidney failure | <15 or dialysis |
* This widely adopted classification required evidence of kidney damage (such as increased proteinuria) above a glomerular filtration rate (GFR) level of 60 mL/min/1.73 m 2 for the diagnosis of CKD but not below this threshold. Chronic kidney disease is defined as either kidney damage or GFR < 60 mL/min/1.73 m 2 for ≥ 3 months. Kidney damage is defined as pathologic abnormalities or markers of damage, including abnormalities in blood or urine test results, or imaging studies.
Since the publication of the NKF CKD guidelines, the most common method used to estimate renal function is the simplified 4-variable Modification of Diet in Renal Disease (MDRD) equation. An important advance in the field has been improved standardization in serum creatinine calibration among different laboratories, specifically the adoption of assays traceable to an isotope-dilution mass spectrometry method. This development is important for epidemiology studies because even relatively small calibration differences in serum creatinine assays can translate into rather large differences for estimated disease prevalence. A newer equation has been proposed and is increasingly being adopted. The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation has the same data input elements (serum creatinine, age, sex, and race) but is derived from a more broad-based set of studies and in general results in lower estimates of CKD prevalence in the general population than the MDRD equation. There are now also equations based on alternate filtration markers such as cystatin C, which can be used alone or in combination with serum creatinine (see Chapter 26 ).
Prevalence
Most of the population epidemiology of CKD has focused on the prevalence of CKD. Probably the best longitudinal data source for determining the prevalence of CKD over time in the population of an entire country has been the National Health and Nutrition Examination Survey (NHANES) in the United States. Sponsored by the U.S. Centers for Disease Control and Prevention (CDC), NHANES is a series of surveys encompassing interviews and physical examinations on a nationally representative sample of participants. The second NHANES (NHANES II) was conducted from 1976 through 1980, and the third NHANES (NHANES III) from 1988 through 1994; the latest NHANES was launched in 1999 as a continuous survey.
Table 20.3 summarizes findings of a number of studies published over the last decade that have examined temporal trends in CKD prevalence in the United States on the basis of NHANES. Most of the studies have shown an increase in the crude prevalence of CKD over time (unadjusted for important secular trends such as aging of the population over time). As Table 20.3 illustrates, differences in choice and calibration of filtration marker and choice of estimating equation result in different estimates of CKD prevalence as well as differences in rates of change of disease burden over time. Regardless of the exact figure, it is clear that the number of individuals with CKD is two orders of magnitude larger than the absolute number of incident ESKD cases, underscoring the public health burden of the disease and how ESKD can truly be considered only the “tip of the iceberg.”
| Study | CKD Definition | Disease Prevalence During Time Period (%) | Change in Prevalence per Year | GFR Estimating Equation | Filtration Marker Calibration and Alignment | ||
|---|---|---|---|---|---|---|---|
| 1976-1980 | 1988-1994 | 1999-2004 | |||||
| Hsu et al, 2004 | eGFRcr < 60 | 2.0 * | 2.5 * | +1.7% per year † | 4-variable MDRD Study Equation | For both time periods, Cleveland Clinic calibrated Cr = Cr − 0.23 | |
| Coresh et al, 2005 | CKD stages 1-4 (eGFRcr and ACR) | 8.8 | 9.4 (1999-2000) | +0.8% per year | 4-variable MDRD Study Equation | 1988-1994: Cleveland Clinic calibrated Cr = Cr − 0.23; 1999-2000: Cleveland Clinic calibrated Cr = Cr + 0.13 | |
| eGFRcr < 60 | 4.4 | 3.8 (1999-2000) | −1.7% per year | ||||
| Coresh et al, 2007 | CKD stages 1-4 (eGFRcr and ACR) | 10.0 | 13.1 | +2.6% per year † | IDMS-traceable MDRD Study Equation | 1998-1994: standardized Cr = −0.184 + 0.960 × Cr; 1999-2000: standardized Cr = 0.147+1.013 × Cr; After 2000: did not require calibration Conservative trend analysis added 0.04 mg/dL to 1988-1994 creatinine values | |
| eGFRcr < 60 (primary analysis) | 5.6 | 8.1 | +3.5% per year † | ||||
| eGFRcr < 60 (conservative trend analysis) | Data not shown | Data not shown | +1.4% per year † | ||||
| Foley et al, 2009 | CKD stages 1-5 (eGFRcyc and ACR) | 15.1 | 14.9 (1999-2002) | −0.1% per year | CKD-EPI cystatin C 2008 | No calibration performed for cystatin C | |
| eGFRcys < 60 | 6.4 | 6.9 (1999-2002) | +0.8% per year | ||||
| Grams et al, 2013 | eGFRcys < 60 | 5.5 | 8.7 (1999-2002) | +4.9% per year † | CKD-EPI Cystatin C 2012 | 1988-1994: standardized cystatin C = 1.12 × [0.022 + 0.80 × (cystatin C)]; 1999-2002: standardized cystatin C = 1.12 × [cystatin C − 0.12] | |
| eGFRcr-cys < 60 | 4.4 | 7.1 (1999-2002) | +5.0% per year † | CKD-EPI Cr-cystatin C 2012 | |||
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
