Residual lifetime risk of chronic kidney disease

Abstract

Background

An accurate estimation of the lifetime risk of chronic kidney disease (CKD) can aid in patient education while also informing the development of public health screening programs and educational campaigns.

Methods

Framingham Offspring Study participants were included if they were free of CKD (estimated glomerular filtration rate <60 mL/min/1.73 m²) at age 50 years and had at least two serum creatinine measures during follow-up (mean 16 years, 49 506 person-years). We estimated the lifetime risk of CKD to age 90 years adjusting for the competing risk of death in the overall cohort and in population subgroups with known CKD risk factors including hypertension, obesity and diabetes.

Results

Overall 3362 individuals (52% women) were included in the study. Mean age at study baseline was 54 years. By the end of the study, 729 individuals (21.7%) developed CKD and 618 (18.4%) died. At age 50 years, the cumulative lifetime risk of CKD was 41.3% [95% confidence interval (CI) 38.5–44.0]. The risk was increased in those with risk factors at baseline including diabetes (52.6%, 95% CI 44.8–60.4), hypertension (50.2%, 95% CI 46.1–54.3) and obesity (46.5%, 95% CI 41.1–52.0). For those individuals without any risk factors at baseline, the lifetime risk of CKD was lower (34.2%, 95% CI 29.4–39.0) relative to those with 1, 2 or 3 risk factors (45.0, 51.5 and 56.1% respectively, P < 0.01 for all compared with those with no risk factors).

Conclusions

Four out of 10 individuals without CKD at age 50 years will eventually develop CKD. This risk is modified by the presence of hypertension, diabetes and obesity at baseline. This demonstrates the importance of early identification of CKD risk factors, to aid in patient education, and potentially to reduce the future risk of disease.

body mass index, chronic renal insufficiency, diabetes mellitus, epidemiology, hypertension

INTRODUCTION

For any individual in a population, the risk of developing a given disease at any particular time is low. Despite this, the lifetime risk of many chronic diseases is high [1–3]. Accurately estimating these risks informs the development of screening programs and public health education campaigns, and can aid in individual patient education by providing a generic risk of developing a disease based on the presence or absence of specific risk factors.

Chronic kidney disease (CKD) is an important public health problem that affects up to 15% of adults in the USA [4, 5]. CKD is a major source of morbidity and is an independent risk factor for both cardiovascular disease (CVD) and all-cause mortality [6–8]. Awareness of CKD in the general population is low, with many individuals unaware that they have this condition [9]. The lifetime risk of end-stage renal disease (ESRD) has been estimated at 1.8–7.3% depending on age and gender based on registry data [10]. However, the overall incidence of ESRD remains low and the majority of individuals with CKD will die of CVD before progressing to this stage. Little is known about the lifetime risk of earlier stages of CKD, which comprise the vast majority of cases in the USA [5]. A recent study estimated the lifetime risk of stage 3 CKD to be 59.1% in US adults [11]. This estimate was based on simulated data extrapolated from prevalence estimates. This could lead to some bias in the estimate due to an inability to adequately account for the competing risk of death. In addition, although hypertension, diabetes and obesity are known key risk factors for the development of CKD [12], the effect of risk factor burden on modifying CKD risk is not well known. Thus, the goal of this study was to estimate the lifetime risk of CKD in individuals from the Framingham Heart Study free of CKD at age 50 years, and to determine how the presence of known CKD risk factors alters lifetime risk of CKD in these individuals.

MATERIALS AND METHODS

Study sample

Participants were drawn from the Framingham Offspring cohort. The Offspring cohort comprises the children and spouses of the original cohort and began enrolling in 1971 [13]. Examinations were done in 4–8 year cycles including a physician interview, physical examination and assessment of CVD factors. Serum creatinine was measured at exam cycle 2 (1979–1983) and cycles 5–9 (1991–2014). For the purposes of this study, individuals were eligible for inclusion if they attended at least two examination cycles between exam 5 and exam 9 and were free of CKD at the baseline examination. Of the 5124 individuals who attended the first examination cycle, 3696 attended exam 5 and at least one more examination cycle. Overall, 270 individuals had CKD at the baseline examination, 14 were missing information on covariates and 24 were aged <50 years at the end of follow-up and were therefore excluded. The final population therefore included 3362 participants. Mean age at study initiation was higher in those excluded from the final analysis (40 versus 34 years, P < 0.001). The excluded participants were more likely to have a history of CVD (31.6 versus 24.7%) and were more likely to die over the course of the study (53.9 versus 17.8%, Supplementary data, Table S1). The study was approved by the Institutional Review Board at Boston University and all participants provided written informed consent.

Outcome definition

Serum creatinine was measured at each examination cycle using the modified Jaffé method. This measurement was calibrated by applying a correction factor of 0.23 mg/dL to National Health and Nutrition Examination Survey III (NHANES III) creatinine values [14]. The creatinine values from this study were then calibrated to the age- and sex-specific means from NHANES III. CKD was defined using the CKD-EPI equation [15] as an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m². Age at CKD diagnosis was defined as the midpoint age between examinations where the participant did and did not have CKD.

Covariate assessment

Risk factors were assessed at the examination cycle attended by the participants closest to age 50 years. Diabetes was defined as a fasting blood glucose of ≥126 mg/dL or diabetes treatment. Systolic and diastolic blood pressures were recorded as the average of two measurements performed by a physician. Hypertension was defined as a systolic blood pressure of ≥140 mmHg or a diastolic blood pressure of ≥90 mmHg, or treatment with anti-hypertensive medications. Body mass index (BMI) was calculated by dividing the weight in kilograms by the square of the height in meters. Obesity was defined as a BMI ≥30 kg/m².

Statistical analysis

To calculate lifetime risk, a modified Kaplan–Meier survival analysis was used [2, 16]. Briefly, person-years follow-up was generated for each participant on the basis of 5-year age grouping beginning at age 50 years. Age-adjusted rates of CKD and death for each 5 year subgroup were estimated using direct standardization. A Kaplan–Meier lifetime risk model was used to estimate the unadjusted cumulative incidence. The adjusted incidence rate was estimated by calculating the conditional probability of both CKD and death. Between-group differences in the adjusted cumulative incidence rates were tested for significance using a mixed linear model for dichotomous outcomes.

Participants were followed to the time of CKD diagnosis, death, age at last contact (if free of CKD) or 90 years. Participants were censored at that point because few individuals live beyond 90 years. Age-specific hazards, incidence ratios, cumulative incidence and survival probabilities were calculated. These estimates were adjusted for the competing risk of death from any cause because traditional censoring at the time of death would tend to overestimate lifetime risk when the competing risk is high [16]. The lifetime risk estimate is the sum of the age-specific incidence of CKD adjusted for mortality from the baseline examination to the last follow-up.

We performed two secondary analyses. First, we stratified the participants according to the presence or absence of specific CKD risk factors at age 50 years (obesity, diabetes and hypertension), and then calculated the overall lifetime risk in the presence or absence of these risk factors. We further calculated the 10-, 20-, 30- and 40-year risk for CKD for each of these strata. Finally, we aggregated risk factors by classifying individuals as having 0, 1, 2 or 3 risk factors at baseline and calculated the lifetime risk and risk per decade of CKD for each of these groups. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC, USA). A two-tailed P-value of <0.05 was considered significant.

RESULTS

Study sample

The baseline characteristics of the participants at the examination closest to age 50 years are shown in Table 1. In total, 3362 individuals (52.5% women) were included in the study. Mean age at entry was 54 ± 9 years and participants were followed for a total of 49 506 person-years. At baseline, 842 individuals (25.0%) were obese, 1101 (32.8%) had hypertension and 246 (7.3%) had diabetes. During the course of the study, 729 participants (21.7%) developed CKD and 618 (18.4%) died.

Table 1

Study sample characteristics at baseline except where otherwise noted

	All	Women	Men
n	3362	1765	1597
Age at entry, years	54.6 (9.4)	54.8 (9.4)	54.5 (9.2)
BMI, kg/m2 (SD)	27.5 (5.0)	28.3 (4.2)	26.7 (5.4)
Obesity, n (%)	842 (25.0)	455 (28.5)	387 (21.9)
Baseline eGFR, mL/min/1.73 m2 (SD)	91.2 (17.0)	90.6 (17.1)	91.7 (16.9)
Hypertension, n (%)	1101 (32.8)	574 (35.9)	527 (29.9)
Diabetes, n (%)	246 (7.3)	147 (9.2)	99 (5.6)
Exam cycle at entry to study
Exam 5 (1987–1991)	2714 (80.7)	1298 (81.3)	1416 (80.2)
Exam 6 (1991–1995)	541 (16.1)	254 (15.9)	287 (16.3)
Exam 7 (1995–1998)	80 (2.38)	32 (2.0)	48 (2.7)
Exam 8 (1998–2001)	27 (0.8)	13 (0.8)	14 (0.8)

Data expressed as mean (SD) for continuous and n (%) for dichotomous variables.

Lifetime risk of CKD

In those individuals without CKD at age 50 years, the overall lifetime risk to age 90 years was 41.3% [95% confidence interval (CI) 38.5–44.0] (Table 2). This risk did not differ significantly between men and women (40.4 versus 42.0%, P = 0.36). The risk was modified by the presence of known CKD risk factors at baseline. In those individuals without hypertension, diabetes or obesity, the lifetime risk of CKD was 34.2% (95% CI 29.4–39.0). The presence of any one of these risk factors at age 50 years was associated with a higher risk of CKD. For those with hypertension, the lifetime risk was 50.2% (95% CI 46.1–54.3), for obesity 46.5% (95% CI 41.1–52.0) and for diabetes 52.6% (95% CI 44.8–60.4) (Table 2).

Table 2

Lifetime risk of CKD Framingham Heart Study participants free of CKD at age 50 years

	N	Cumulative lifetime risk of CKD (%)	95% CI	P-value
Overall	3362	41.3	38.5–44.0
Sex
Women	1765	42.0	37.9–46.2
Men	1597	40.4	36.7–44.0	0.55
Hypertension
Yes	1101	50.2	46.1–54.3
No	2261	36.0	31.9–40.2	<0.001
Diabetes
Yes	246	52.6	44.8–60.4
No	3116	40.4	37.4–43.4	0.004
Obese
Yes	842	46.5	41.1–52.0
No	2520	39.4	36.2–42.6	0.03
Number of risk factors
0	1793	34.2	29.4–39.0
1	1030	45.0	40.6–49.4	0.001
2	458	51.5	45.1–57.8	<0.001
3	81	56.1	41.5–70.6	0.005

P-values reflect the comparison between individuals with and without risk factors.

Cumulative risk of CKD stratified by the presence of risk factors at baseline

Figure 1 shows the cumulative risk of CKD per decade from ages 50 to 90 years adjusted for the competing risk of death from any cause in individuals with and without hypertension, obesity and diabetes. In all groups, except those with diabetes at baseline, the 10-year risk of CKD was low but increased rapidly after age 70 years. This is expected given the known increased prevalence of CKD in older age groups. In individuals with diabetes, the risk increased rapidly from the time of entry into the study. The overall lifetime risk of CKD in participants with diabetes was not markedly higher than those with hypertension, likely because of an increased risk of early mortality in the diabetes group (Table 3). At all time points, the presence of any one of these risk factors was associated with a higher risk of CKD. For example, the 20-year risk of CKD was higher in individuals with hypertension (26.2%), diabetes (34.0%) and obesity (20.9%) relative to those with no risk factors at the time of study entry (13.0%).

Table 3

Cumulative lifetime risk of incident CKD stratified by the presence of CKD risk factors

		Cumulative risk of CKD (%)
	n	10 year	95% CI	20 year	95% CI	30 year	95% CI	40 year	95% CI
Overall	3622	4.0	3.1–5.0	17.4	15.8–19.0	33.9	31.7–36.1	41.3	38.5–44.0
Number of risk factors
0	1793	3.1	2.1–4.1	13.0	11.0–14.9	25.6	22.5–28.6	34.2	29.4–39.0
1	1030	4.5	2.5–6.5	19.7	16.6–22.9	39.4	35.6–43.3	45.0	40.6–49.4
2	458	6.0	2.3–9.7	26.4	21.1–31.8	43.6	37.7–49.6	51.5	45.1–57.8
3	81	23.7	6.6–40.7	41.9	26.2–57.7	50.1	36.1–65.8	56.1	51.5–70.6
Risk factor
Hypertension	1101	7.4	4.5–10.3	26.2	22.6–30.0	43.7	29.8–47.6	50.2	46.1–54.3
Obesity	842	5.5	3.3–7.7	20.9	17.5–24.2	37.9	33.5–42.3	46.5	41.1–52.0
Diabetes	246	11.5	4.3–18.7	34.0	25.9–42.1	49.1	41.3–56.9	52.6	44.8–60.4

FIGURE 1

Cumulative risk of CKD at time points after the baseline examination in the entire cohort and in individuals with and without hypertension, diabetes and obesity.

Figure 2 shows the cumulative risk of CKD in individuals stratified by the number of risk factors at baseline. In total, 1793 participants (53.3%) did not have hypertension or diabetes and were not obese. A total of 1030 (28.4%) had at least 1 risk factor, 458 (12.6%) had 2 risk factors and 81 (2.4%) had all 3 risk factors. The distribution of baseline characteristics stratified by the number of risk factors is shown in Supplementary data, Table S2. In individuals without any of these three risk factors at baseline, the lifetime risk of CKD was low 34.2% (95% CI 29.4–39.0). There was a graded increase in the risk of CKD with increasing numbers of risk factors at baseline. Individuals with 1, 2 and 3 risk factors had a cumulative lifetime risk of 45.0% (95% CI 40.6–49.4), 51.5% (95% CI 45.1–57.8) and 56.1 (95% CI 51.5–70.6), respectively (P < 0.005 for all groups compared with individuals with no risk factors).

FIGURE 2

Cumulative risk of CKD at time points after the baseline examination stratified by number of risk factors present at baseline. Risk factors include hypertension, diabetes and obesity.

DISCUSSION

The findings of this study are 4-fold. First, ∼4 in 10 individuals free of CKD at age 50 years will go on to develop CKD in their lifetimes. Second, the lifetime risk of CKD is modified by the presence of diabetes, obesity and hypertension at baseline. Third, individuals without any of these risk factors face a considerably lower long-term risk of CKD. Finally, there is a graded increase in CKD risk with increasing numbers of risk factors at baseline.

In the context of the current literature

In this study, we found that the lifetime risk of stage 3 CKD for an individual free of CKD at age 50 years was 41.3%. This estimate is considerably lower than that reported in a prior study of US adults, which estimated the lifetime risk of CKD stage 3 at birth at 59.1% [11]. In that study, the risk was higher in women and African Americans. Another recently published simulation study, also based on NHANES, found that the residual lifetime risk of CKD in US adults aged 50–64 was 52% [17]. There are a number of potential reasons for why the estimate was lower in our cohort. First, these estimates of lifetime risk were done by simulating outcomes based on population prevalence estimates. Because each individual was not followed up, this may not have accounted sufficiently for the competing risk of death from any cause. Traditional survival analysis treats death as a censoring event. This assumes that the risk of the disease is identical in the censored group as in those who continue to be followed. However, a person who dies cannot develop CKD and as a result, their risk for subsequent events is zero. Not accounting for the competing risk of death can lead to a significant overestimation of the lifetime risk. This is particularly the case with CKD where the majority of cases occur beyond the age of 70 years where the competing risk of death is high.

A second potential explanation for the lower lifetime risk estimate in this study is that the Framingham Heart Study Offspring cohort comprises a population of individuals of European-American ancestry. The absence of African Americans from this study would tend to lower the calculated risk because of the considerably higher prevalence of CKD in that population [5]. Finally, more than one-third of Offspring cohort participants were not included in this study. The majority were excluded because they did not attend two examinations beyond exam 4. Individuals excluded for this reason were older and were more likely to have died or have had CVD. The exclusion of these participants may have biased our estimate downwards as they had a higher prevalence of CKD risk factors at baseline.

More recently, a study using a prospective cohort of individuals in Iceland found that the overall lifetime risk of CKD in that population was 35.8% in women and 21% in men [18]. While this is closer to the lifetime risk estimate found in our study, the marked disparity between men and women is difficult to explain. The authors conducted a number of sensitivity analyses but the difference persisted no matter which definition of CKD was used. An excess mortality in men could have led to a difference in the lifetime risk estimate but it seems unlikely that it would be accounted for by this alone. In contrast, in our study, we found that the lifetime risk of CKD was similar in women and men. This is similar to the results seen in the prior simulation study using data from NHANES [17].

Implications

In this study, the risk of CKD was modified by the presence of risk factors in middle age with individuals carrying even one risk factor having a higher risk of CKD up to 40 years later. There was a graduated increase in risk at all time-points with the addition of more risk factors. In contrast, individuals with no risk factors at baseline had a relatively low risk of developing future CKD. The fact that the presence of risk factors at age 50 years modified the future risk of CKD up to 40 years later has important implications for the development of population-based strategies to lower the burden of CKD. Risk factor modification near the time of diagnosis of CKD may be too late because it takes place after potential long-term exposure to these risk factors. For example, elevated risk factors in early adulthood are associated with future CVD events [1, 19], while individuals who have few or no risk factors at an early age have a lower lifetime risk of CVD [20]. Similarly, individuals who later develop CKD have elevated risk factors up to 30 years prior to the diagnosis [21]. Thus, risk factor modification at or near the time of CKD diagnosis takes place in the context of long-term exposure to these risk factors during which time significant renal function decline has already occurred. It is possible that earlier identification and management of these risk factors may be more successful at preventing CKD onset. This could aid in patient education, guide public policy decisions and potentially reduce the future prevalence of kidney disease.

Strengths and limitations

There are a number of important strengths of this study including the well-characterized Framingham Heart Study cohort with a long duration of follow-up and multiple measures of renal function. In contrast to simulation studies, we were able to definitively account for the competing risk of death. Finally, we were able to generate estimates of future risk of CKD based on the presence or absence of known CKD risk factors. However, there are several limitations that warrant mention. First, the study population was entirely European-American and as a result, these estimates cannot be extrapolated to other ethnic groups where the risk of CKD may be higher. Participants who were older and sicker at the time of enrollment were more likely to miss follow-up Framingham Heart Study examinations; given that their burden of CKD risk factors was higher than those who remained in the study, they may have been more likely to develop CKD. Finally, because of the low incidence of ESRD in this cohort, we were unable to estimate the lifetime risk of dialysis.

Conclusion

The lifetime risk of CKD in the general population is high, and 4 out of 10 individuals without CKD at age 50 years will eventually develop CKD. The cumulative risk increases rapidly in older age and is modified by the presence of hypertension, diabetes and obesity in mid-life.

SUPPLEMENTARY DATA

Supplementary data are available online at http://ndt.oxfordjournals.org.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to declare. The results presented in this paper have not been published previously in whole or in part.