Familial Risks of Kidney Failure in Sweden: A Nationwide Family Study

Background The value of family history as a risk factor for kidney failure has not been determined in a nationwide setting. Aim This nationwide family study aimed to determine familial risks for kidney failure in Sweden. Methods The Swedish multi-generation register on 0–78-year-old subjects were linked to the Swedish patient register and the Cause of death register for 1987–2010. Individuals diagnosed with acute kidney failure (n = 10063), chronic kidney failure (n = 18668), or unspecified kidney failure (n = 3731) were included. Kidney failure patients with cystic kidney disease, congenital kidney and urinary tract malformations, urolithiasis, and rare inherited kidney syndromes, and hyperoxaluria were excluded. Standardized incidence ratios (SIRs) were calculated for individuals whose parents/siblings were diagnosed with kidney failure compared to those whose parents or siblings were not. Results The concordant (same disease) familial risks (sibling/parent history) were increased for chronic kidney failure SIR = 2.02 (95% confidence interval, CI 1.90–2.14) but not for acute kidney failure SIR = 1.08 (95% CI 0.94–1.22) and for unspecified kidney failure SIR = 1.25 (95% CI 0.94–1.63). However, the discordant (different disease) familial risk for acute kidney failure SIR = 1.19 (95% CI 1.06–1.32) and unspecified kidney failure SIR = 1.63 (95% CI 1.40–1.90) was significantly increased in individuals with a family history of chronic kidney failure. The familial risk for chronic kidney failure was similar for males SIR = 2.04 (95% CI 1.90–2.20) and females SIR = 1.97 (95% CI 1.78–2.17). Familial risks for chronic kidney failure were highest at age of 10–19 years SIR = 6.33 (95% CI 4.16–9.22). Conclusions The present study shows that family history is an important risk factor for chronic kidney failure but to a lower degree for acute kidney failure and unspecified kidney failure.


Introduction
Chronic kidney disease (CKD) is a worldwide medical problem with poor outcomes, high costs, and increased risk of cardiovascular comorbidities and allcause mortality [1][2][3]. In developed countries, it is associated with old age, diabetes, hypertension, obesity, and cardiovascular disease [1]. Diabetic glomerulosclerosis and hypertensive nephrosclerosis are the presumed pathological entities but exact diagnosis is often difficult [1]. Familial and genetic factors are increasingly recognized as important for the development of CKD and endstage renal disease (ESRD) [4][5][6][7][8]. Ferguson et al. first reported that family history of ESRD is common among African Americans with ESRD [9]. Several case-series and case-control studies have confirmed the importance of family history of kidney disease in different populations of patients with CKD and/or ERSD [10][11][12][13][14][15][16][17][18]. However, few follow-up studies have determined the importance of family history of CKD and/or ESRD [19]. In one such study, Hsu et al. found a modest effect of self-reported family history of kidney disease (hazard ratio (HR)51.40) [19]. No study has determined whether familial factors influence the risk of acute kidney failure, which is an increasing global problem [20].
Though multiple genetic loci have been associated with progressive kidney failure and function, [21][22][23] heritability estimates suggests that only a small proportion of the total heritable contribution to the phenotypic variation of CKD have been identified. Large-scale follow-up studies to determine the importance of family history of CKD may therefore be of clinical value for risk assessment, and may help for planning of genetic studies. Clustering of a disease in families may be caused both by both genetic and non-genetic factors [24]. Increased familial risks may indicate shared environmental and lifestyle factors are of importance for disease development, and not only inherited biological factors [25]. However, without familial clustering a genetic cause is unlikely [24].
To our knowledge, there has not been any nationwide follow-up study whose aim was to determine familial risks of kidney failure among offspring/siblings. This nationwide follow-up study determined the familial risks of different forms of kidney failure -chronic kidney failure, acute kidney failure and unspecified kidney failure (i.e., not specified whether it is acute or chronic) -in the offspring/ siblings of individuals with kidney failure. The present study underlines the importance of familial factors in kidney failure.

Materials and Methods
The dataset used in this study was constructed by linking several national Swedish registers provided by the Swedish government-owned statistics bureau Statistics Sweden and the National Board of Health and Welfare [26]. The Swedish multigenerational register contains information on family relationships for index persons born in Sweden in 1932 and later. Individuals born in 1932 or later and who were alive 1987 constituted the present study population. Linkages were made to National Census data in order to ascertain individual-level socioeconomic status, to the Swedish cause of death register , to the Swedish outpatient care register (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010), and to the Swedish hospital discharge register , the last of which records nationwide dates of hospitalization and hospital diagnoses since 1987. All linkages were performed using the individual national identification number that is assigned to each resident in Sweden for their lifetime. This number was replaced by a serial number in order to preserve anonymity. The serial numbers were used to check that each individual was entered only once (for his or her first main or secondary diagnosis of kidney failure). Over 8.1 million individuals and their biological parents (3.8 million families) were included in the database; the oldest (born in 1932) were 78 years at the end of follow-up period, which ran from 1987-2010.

Statistical Analysis
For the analysis of familial risks of kidney failure, a previously described method was used [27]. The method is described in detail by Hemminki et al [28] and takes into account clustering within families, since it is based on complete ascertainment of sib ships in affected individuals. Person-years at risk (i.e., the number of persons at risk multiplied by the time at risk) were calculated from the start of the follow-up on 1 January 1987 until diagnosis for kidney failure, death, emigration, or the end of the follow-up (31 December 2010) [29]. Age-adjusted incidence rates were calculated for the whole follow-up period, divided into 5-year periods [29]. Standardized incidence ratios (SIRs) were used to measure the relative risk of kidney failure in individuals with one or more parents with a history of kidney failure compared with individuals with parents without a history of kidney failure. Similar calculations were performed separately for siblings.
The familial SIRs were calculated as the ratio of observed (O) and expected (E) numbers of kidney failure cases using the indirect standardization method: where O~P o j denotes the total observed number of cases in the study group; E * (the expected number of cases) is calculated by applying stratum-specific standard incidence rates (l * j ) obtained from the reference group to the stratum-specific person-years of risk (n j ) for the study group; o j represents the observed number of cases that the cohort subjects contribute to the jth stratum; and J represents the strata defined by cross-classification of the following adjustment variables: age (5-year groups), sex, socioeconomic status, time period (5-year groups), geographic region of residence, and comorbidities. 95% confidence intervals (95% CIs) were calculated assuming a Poisson distribution [29].
Data values are accurate to two decimals places. All analyses were performed using SAS version 9.2 (Institute, Cary, NC, USA).

Ethical Considerations
Statistics Sweden and the National Board of Health and Welfare maintain the nationwide registers used in the present study. This study was approved by the Ethics Committee at Lund University (approval number 409/2008 Lund with complementary approvals dated September 1, 2009, and January 22, 2010) and recommendations of the Declaration of Helsinki were complied with. The ethics committee waived informed consent as a requirement.

Results
We analyzed familial risks of kidney failure in the siblings/offspring (aged 0-78 years) of individuals with kidney failure between 1987 and 2010 in Sweden. The population and number of diagnosis for kidney failure are presented in Table 1. A total of 8054071 individuals were included in this cohort. A total of 32462 individuals were diagnosed with kidney failure, 64% (20688) were males and 36% (11774) females (Table 1). Of these patients, 31.0% (10063) were diagnosed with acute kidney failure, 57.5% (18668) with chronic kidney failure, and 11.5% (3731) with unspecified kidney failure. Comorbidities were more common in patients with kidney failure than in the general population ( Table 1). The lowest incidence rates for kidney failure were observed for children ( Figure 1). The incidence rate for kidney failure increased with age in both sexes (Figure 1). At older ages, the incidence rate for kidney failure was higher for males than females ( Figure 1). The incidence rate was highest for chronic kidney failure, and lowest for unspecified kidney failure ( Figure 2).

Familial risk of kidney failure
Familial risks of kidney failure according to disease subtypes are presented in Table 2. Familial risks were adjusted for age, sex, time period, region of residence, socioeconomic status, and comorbidities. The incidence rates for familial and non-familial kidney failure are presented Figure 3. Concordant (same disease in proband and exposed relative) and discordant (different disease in proband and exposed relative) risks were determined. The familial risks were highest for chronic kidney failure: the concordant familial SIR for chronic kidney failure was 2.02. The concordant familial risk was not significantly increased for acute kidney failure (SIR51.08) and for unspecified kidney failure (SIR51.25) ( Table 2). However, discordant risks show that family history (sibling/parent) of chronic kidney failure is a risk factor for both acute kidney failure (SIR51. 19) and unspecific kidney failure (SIR51.63) ( Table 2). Moreover, discordant risks show that family history (sibling/parent) of acute kidney failure is a risk factor for both chronic kidney failure (SIR51.10) and unspecific kidney failure (SIR51.30) ( Table 2). Family history of unspecified kidney failure (sibling/parent) was a risk factor for chronic kidney failure (SIR51.31) ( Table 2). Family history of all kidney failure was a risk factor for all types of kidney failure (Table 2). Familial risks of kidney failure were determined in both males and females. There were no major sex differences (Table 2). In Table 3, familial concordant risks are presented according to the affected relative. Familial risks were adjusted for age, sex, time period, region of residence, socioeconomic status, and comorbidities. Sibling history of chronic kidney failure showed the highest familial risk, with a concordant SIR of 2.52. The familial concordant risk for individuals with a parental history of chronic kidney failure was 1.67. There were no major sex differences. The familial concordant risks for acute and unspecified kidney failure were not significant ( Table 3).
The familial concordant risks (parent/sibling history) were stratified according to age at diagnosis ( Table 4). The familial risks for chronic kidney failure were highly age dependent and were highest risks at younger ages (SIR56.33 between the age of 10 and 19 years). Increased concordant familial risk of 1.81 was noted also for chronic kidney failure for those aged 60 years or more ( Table 4). The familial concordant risks for chronic kidney failure were increased in all age groups except those younger than 10 years. For acute kidney failure, the familial concordant risks were only significantly increased only in two age groups ( Table 4). The familial risk for acute kidney failure before age of 10 years was high (SIR514.21). The age of these six children with familial acute kidney failure were 0, 1, 1, 5, 5, and 7 years. For three children, the diagnosis was unknown (two had ICD diagnosis5Z038 and one had no additional diagnosis). One child was prematurely born (,28 weeks) and/or had a very low birth weight (,1000 g) (ICD-95765A), one had unspecified infectious gastroenteritis (ICD-95009B), and one had gastroenteritis with Escherichia coli (ICD-95008A). No significant increased risk for unspecified kidney failure was observed for any other age groups. However, the familial risk for all kidney failure was increased in all age groups (Table 4).

Test for the extent of the shared non-genetic familial contribution
In order to test for the extent of environmental sharing in the observed risks of kidney failure SIRs for siblings according to difference in age were calculated (Table S1). Familial risks were adjusted for age, sex, time period, region of residence, socioeconomic status, and comorbidities. Overall, the age difference

Additional analyses
In Table S2, familial concordant and discordant risks are presented according to the affected relative. Familial risks were adjusted for age, sex, time period, region of residence, socioeconomic status, and comorbidities. The results were basically similar to the familial concordant/discordant risk in Table 2. Thus, concordant and discordant risk was generally highest for chronic kidney failure, followed by unspecified kidney failure, and weakest for acute kidney failure independent of the type of affected relative (sibling/parent, parents, mother, father or sibling).  Table S3 shows age stratified concordant and discordant familial risks (parent/ siblings) of kidney failure. Familial risks were adjusted for age, sex, time period, region of residence, socioeconomic status, and comorbidities. The results were basically similar to the familial age stratified concordant risks in Table 4. Thus, age stratified concordant and discordant risks were generally highest for chronic kidney failure, followed by unspecified kidney failure, and weakest for acute kidney failure independent of the type of affected relative (sibling/parent, parents, mother, father or sibling). However, for acute kidney failure, the familial concordant risks were highly increased in the two youngest age groups (Table S3). Table S4 presents concordant and discordant familial risks (parent/siblings) after exclusion of patients with kidney cancer in parents/offspring. This did not change the results to any major degree.  Table S5 shows concordant and discordant familial risks (parent/siblings) for the follow up period 2001-2010. The familial risks were similar compared to using a follow-up period from 1987-2010. Thus, inclusion of outpatients with kidney

Discussion
The present study is the first nationwide follow-up study to evaluate the familial risks of chronic, acute, and unspecified kidney failure among offspring/siblings of affected individuals. The results confirm previous case-series and case-control studies, which showed that familial factors are important for chronic kidney failure [9][10][11][12][13][14][15][16][17][18]. The present study adds follow-up data for a whole country. Previously a follow-up study only showed moderately increased familial risk (HR51.40) [19]. The present results indicate that familial factors are important for chronic kidney failure in both males and females of all ages (except ,10 years), although the familial risks were highest at ages 10-19 years (Table 4). This is in contrast to the findings that most risk alleles from genome wide association studies add little to the prediction of CKD [21][22][23]. It is possible that there are a large number of risk alleles that have yet to be discovered that may account for this discrepancy. Unique familial (environmental or genetic) factors may predispose individuals to chronic kidney failure. Support for a genetic contribution to chronic kidney failure comes from the observation that age difference between siblings had little influence (Table S1). If environmental factors were strong, one would expect higher risks for siblings with smaller age differences. For chronic kidney failure, the familial concordant risks were high ( Table 2). The familial concordant risks for acute and unspecified kidney failure were not significant (Table 2). However, increased discordant familial risks show that familial factors also are involved in acute and unspecified kidney failure (Table 2), though to a lower degree than for chronic kidney failure. Acute kidney failure is instead more related to precipitating factors such as sepsis, complex surgery, diagnostic procedures requiring intravenous contrast continue, and druginduced kidney injury [20]. Unspecified kidney failure is probably a mixture of patients with chronic and acute kidney failure. No previous study has reported familial risks for acute and unspecified kidney failure. Of interest is the high familial risk for acute renal failure among children younger than ten years (SIR514.21) and also children and teenagers between 10 and 19 years of age (SIR52.52) ( Table 4). Among the children younger than 10 years, 2 were related to infection and one to preterm birth and or/low birth weight, which argue against a genetic cause for these cases of acute renal failure. Three cases were unknown and we cannot exclude that in rare cases familial factors are important among children and teenagers. A study from Norway of acute renal failure identified 315 cases of acute renal failure among children under the age of 16 years [30]. The estimated incidence rate was 3.3 cases per 100 000 children. This is in range of out overall kidney failure incidence rate among children (Table 1). Most cases (43%) in the Norwegian study were children under the age of five years [30]. The authors identified 53 aetiologies and classified these into 30 aetiological groups: 25% were prerenal failure (n575), 74% were intrinsic/ renal failure (n5234), and 2% were postrenal failure (n55). Nephritic syndromes was the most common cause (44%) of acute kidney failure, followed by haemolytic-uraemic syndrome (HUS) (15%) [30].
The present design has potential advantages and disadvantages. Strengths of the study include complete nationwide coverage from 1987 in a country with high standards of diagnosis, with diagnoses often being made by specialists during extended examinations in clinics. The Swedish hospital discharge register contains no information about diagnostic procedures, which is a limitation. Moreover, the validity of ICD codes for kidney disease has not been reported. However, the Swedish hospital discharge register has been extensively validated and its overall diagnostic validity is close to 90% [31][32]. A limitation is the inclusion of asymptomatic early stages of renal failure. The Swedish ICD-9 code 585 has no sub codes for different stages of ESRD. Thus, a number of patients with unidentified early stages of renal failure are not included in the study, which most likely is a non-differential bias with regards to familial risks. Another likely nondifferential bias regarding familial risks is that cases in probands and relatives before 1987 are unknown. Moreover, the number of comorbidities is rather low ( Table 1), possible due to that diagnosis made in primary health are not included. No nationwide primary health care register exists in Sweden.
Another important strength of our study is that it was based on nationwide registers and was thus free of selection and recall bias. The Swedish multigeneration register and the Swedish hospital discharge register are validated data sources that have been proven to be reliable in the study of many diseases [26], [31]. Data in our dataset are almost 100% complete [26].
In summary, the present study found indications of strong aggregation of chronic kidney failure, while familial factors are less important in acute and unspecified kidney failure. Familial non-genetic factors contribute among husbands but not wives. Identification of the unique familial factors in chronic kidney failure will advance our knowledge about the pathogenesis of kidney failure.