Ethnicity, Socioeconomic Status, and Attainment of Clinical Practice Guideline Standards in Dialysis Patients in the United Kingdom

Abstract

Background and objectives: The role of socioeconomic status (SES) and its contribution to ethnic differences in standards attainment among dialysis patients is not known.

Design, setting, participants, & measurements: We examined associations between area- level SES (Townsend index) and ethnicity (white, black, South Asian) and standards attainment in 14,117 incident dialysis patients (1997–2004) in the UK.

Results: Deprived patients were less likely to achieve hemoglobin (Hb) ≥ 10g/dl (trend P < 0.001) but not after controlling for patient and center characteristics (trend P = 0.1). There was no association with hemodialysis dose and parathyroid hormone (PTH) standard but deprived patients had better attainment of phosphate (PO4) <5.6 mg/dl, calcium (Ca) and Calcium-phosphate (CaPO4) standard (e.g., most deprived versus least deprived adjusted odds ratio [OR] 1.25, 95% confidence intervals [CI] 1.12, 1.38). There was no association with SES using a lower limit for PO4 (3.5 − 5.5 mg/dl). Compared with Whites, Blacks had lower attainment of Hb (adjusted OR 0.57, 95% CI 0.45, 0.71) and PTH standards (adjusted OR 0.27, 95% CI 0.22, 0.33) but better attainment of PO4 and CaPO4, while South Asians experienced better or comparable outcomes for most standards except Ca and PTH.

Conclusions: There was no evidence of socioeconomic inequity in standards attainment or a consistent pattern of inequity by ethnic group. The lower attainment of some standards in ethnic minorities may reflect biologic differences rather than ethnicity-related inequity of care.

Assuring quality of care for patients with kidney disease has been central to initiatives such as K/DOQI (1)and the UK Renal Association guidelines (2), and is a key role of national and international renal registries. One important aspect of quality assurance is to assess and minimize any inequities in the health care received by key social groups (e.g., socioeconomic status [SES], ethnic minorities) to reduce inequities of patient outcome.

Socioeconomic and ethnic disparities in indicators of health outcome have been reported among patients on renal replacement therapy (RRT). Socially deprived patients in the US have lower hemoglobin (Hb) (3) and lower rates of erythropoietin (EPO) use at the start of RRT (4). Once on dialysis in the US, Blacks tend to achieve lower Hb levels (5), lower dialysis dose (5), and higher parathyroid hormone (PTH) levels (6) than Whites, whereas patients of South Asian descent achieve Hb levels comparable to those of Whites and a higher dialysis dose (7). The extent to which such ethnic disparities observed in the US are explained by socioeconomic differences or are due to ethnicity per se remains unclear.

Ethnic disparities in RRT-related outcomes have mostly been described in the US where socioeconomic factors and variation in access to health care may explain some of the disparities. Where health care is publicly funded through taxation or social insurance, there is some evidence that RRT-related outcomes vary less between social deprivation groups. In England and Wales, for example, survival on RRT has been shown to be comparable in patients regardless of area-level deprivation (8). And in the transplant setting, in contrast to studies from the US (9–11), a study of kidney transplant survival in France has demonstrated similar outcomes in White and Black patients (12).

The situation can be complex if the association between SES and a health outcome differs between ethnic groups, in other words, there is an interaction between SES and ethnicity. In the US dialysis setting, such an interaction has been described for mortality on dialysis, with low income only associated with increased mortality among Black patients (13). Interactions between SES and ethnicity have also been reported in nonrenal chronic diseases, such as cardiovascular disease and diabetes (14–15), cancer mortality (16), alcohol dependence (17), and bronchial asthma (18).

The aim of this study was to examine the association between SES, ethnicity, and any potential interactions, and the attainment of national clinical practice guideline standards in dialysis patients in the UK, where the health care system is funded by general taxation and free at point of care.

Materials and Methods

Study Population

The UK Renal Registry (UKRR) collects clinical and biochemical data for all patients receiving RRT in the UK, and the data collection methods have been described in detail elsewhere (19). Only dialysis centers in England and Wales (E&W) were routinely submitting quarterly laboratory data at the time of the study and could be included. All incident adult (>18 yr) patients starting RRT between January 1, 1997, and December 31, 2004, and alive on either hemodialysis (HD) or peritoneal dialysis (PD) at one year from start of RRT were considered. For patients with missing ethnicity data on the UKRR database, ethnicity was updated in a hierarchical order: from the UK Transplant database, the independent organization responsible for maintaining the national organ donor register in the UK, using the name-recognition software program SANGRA, developed and validated in the UK to identify South Asian ethnic origin by name (20); and finally by recoding patients with missing ethnicity as Whites if they lived in a UK Census 2001 ward that had a predominant White population (>98%). The incidence of RRT is three to four times higher in ethnic minorities compared with Whites in the UK (21,22), and therefore updating ethnicity using Census statistics could potentially lead to a misclassification error of up to 6% to 8%, which was considered acceptable. Patients whose ethnicity could not be updated by the above methods and those with missing primary renal disease and residence postcode were excluded from the study.

Measurement of Social Deprivation

The UKRR does not collect individual-level SES data. Therefore, area-level SES measured using the Townsend Index was used as an ecological proxy for individual-level SES. This score was derived for each 2001 UK Census output area and based on the percentage of unemployed individuals and the percentage of households that had no car, were overcrowded, and were not owner-occupied (23). Each patient's postcode of residence was matched to the 2001 UK Census output area file. The postcodes were then divided into five equal-sized population quintiles according to the level of deprivation of the area in which they were in, a high Townsend score indicating more deprivation. For the 5% of postcodes that cross an output-area boundary and therefore have more than one Townsend score, the mean value was taken.

Measurement of Clinical Practice Guideline Standards

The standards studied were Hb, serum phosphate (PO4), calcium (Ca), and PTH on all dialysis patients and hemodialysis dose (measured as urea reduction ratio [URR]) for patients on HD. The UKRR does not collect data on dialysis dose for PD patients. Laboratory values from the fourth quarter of the first year of RRT (or third quarter if a fourth-quarter reading was not available) were used to ascertain achievement of standards set by the UK Renal Association (2): Hb ≥ 10 g/dl; PO4 < 1.8 mmol/L (<5.6 mg/dl); Ca 2.2 to 2.6 mmol/L (8.8 to 10.4 mg/dl); PTH <4 times the upper limit of normal range ∼ <32 pmol/L (<300 pg/ml); URR >65%. There were no standards set for calcium phosphate product (CaPO4) by the UK Renal Association and therefore the KDOQI target of <55 mg²/dl² was adopted. For patients on HD, all the variables were measured predialysis. Patients who did not have a recorded value for a laboratory variable either in the third or fourth quarter in the first year of RRT were excluded from the analysis for that laboratory standard. For the analysis of URR, patients on home dialysis and those receiving dialysis less than three times per week, were excluded.

Statistical Analyses

To overcome the complex links between SES and ethnicity, we first examined the role of SES on attainment of standards in an ethnically homogenous subgroup of incident White dialysis patients (n = 12,411). The association between ethnicity and attainment of standards is then examined in all patients (n = 14,117) adjusting for patient-level factors, SES, and center effects.

The independent variables of interest were ethnicity and SES. Patients were grouped into three ethnic groups: White, Black, and South Asian (Indian subcontinent). Patients of Chinese and other mixed origin were excluded, as they were a heterogeneous group and accounted for only a small proportion of patients (1.7%). Chi-squared test and Kruskal Wallis test were used to compare baseline characteristics. Separate multivariable logistic regression models were used to study the independent effect of SES and ethnicity on attainment of UK Renal Association standards for each of the laboratory variables.

The covariates sequentially included in the multivariable logistic regression analyses were age group (18 to 34, 35 to 44, 45 to 54, 55 to 64, 65 to 74, ≥75 yr), gender, primary renal disease, year of start of RRT, RRT modality, and center. It was hypothesized that age group, gender, year of start of RRT, and center may act as confounders, but the other variables above may be intermediaries that mediate any association. We included center variables to adjust for variations in clinical management, which may confound geographically patterned sociodemographic factors. In the UK, there is considerable variation in both the mean Townsend score and ethnic composition for each center (24) and in the proportion of patients in each center achieving clinical practice guideline standards (25).

We calculated odds ratios, 95% confidence intervals, and p-values using robust standard errors, as patients are clustered within centers and hence conventional standard errors need to be inflated by the design effect. As there were 49 centers, we undertook interaction tests between SES and center by grouping centers into tertiles based on the percentage of patients in each center achieving the standard. The effect of interaction parameter was determined using Wald tests, as likelihood ratio tests are not appropriate when robust standard errors are used to account for clustering. The association of ethnicity and attainment of standards was then examined by including all patients and using the same models as above, but also adjusting for area-level SES before center adjustment.

Sensitivity Analyses

Updating ethnicity with SANGRA and census statistics could have resulted in residual confounding due to misclassification error. However, results were similar when we repeated the analyses including only patients whose ethnicity was available on the UKRR database (n = 12171). Therefore, only results using the ethnicity updated cohort are shown.

All statistical analyses were performed using SAS software version 9.1.

Results

Study Cohort

Of the 16,050 patients who started RRT during the study period and were alive either on HD or PD at 1 yr from start of RRT, 1681 (10.4%) patients were excluded either due to missing cause of end-stage renal disease (ESRD) (n = 501), Townsend score (n = 250), or missing ethnicity (n = 930). Patients of Chinese and mixed ethnic origin (n = 252) were excluded, as they represented a heterogeneous group, resulting in a final study cohort of 14,117 patients. Of these, the data completeness for each of the laboratory variables was: Hb (92%), URR (80%), PO4 (92%), Ca (92%), PTH (74%). Patients who did not have a recorded value for a laboratory variable either in the third or fourth quarter in the first year of RRT were excluded from the analysis for that laboratory standard. The age and gender distributions, frequency of diabetes, Townsend score, and ethnicity among patients with and without data for any of the laboratory variables were comparable (data not shown).

Baseline Characteristics

The study cohort had a median age of 62.8 yr (interquartile range 49.4 to 72.3 yr); 62.0% were males and 86.4% were White. Diabetes was the cause of ESRD in 20.2%. The baseline characteristics of White patients according to SES and of all patients according to ethnic group are shown in Table 1 and Table 2, respectively.

Table 1.

Baseline characteristics of White patients according to area-level social deprivation

	Quintile 1	Quintile 2	Quintile 3	Quintile 4	Quintile 5	Total	P value
Number of patients (%)	2307 (18.6)	2628 (21.2)	2505 (20.2)	2637 (21.2)	2334 (18.8)	12,411	–
Median age (interquartile range) years	65.2 (53.2, 74.2)	65.0 (53.1, 73.7)	64.2 (51.4, 73.1)	62.9 (48.7, 72.9)	60.8 (46.1, 71.7)	63.7 (50.5, 73.1)	<0.0001
Males (%)	64.5	63.2	63.4	62.2	60.4	62.7	0.04
Treatment modality
Haemodialysis (%)	58.6	62.3	65.8	68.9	74.1	66.0	<0.0001
Cause of renal failure (%)							<0.0001
Uncertain	24.5	26.5	26.6	23.7	23.2	24.9	–
Diabetes	13.9	15.6	17.8	20.2	22.0	17.9	–
Renovascular disease	13.8	13.4	13.5	13.8	15.3	13.9	–
Other	16.3	14.8	14.0	13.2	12.0	14.0	–
Glomerulonephritis	13.1	11.9	11.1	12.0	10.3	11.7	–
Polycystic kidney disease	9.6	8.6	8.1	7.9	12.0	8.2	–
Chronic pyelonephritis	8.9	9.3	8.9	9.3	10.5	9.4	–

Table 2.

Baseline characteristics of all patients according to ethnicity

	White	South Asian	Black	All patients	P value
Number of patients (%)	12,411 (87.9)	1153 (8.2)	553 (3.9)	14,117	–
Median age (interquartile range) years	63.7 (50.5, 73.1)	58.5 (47.3, 66.6)	54.1 (38.5, 67.0)	62.8 (49.4, 72.3)	0.0001
Males (%)	62.7	59.2	52.1	62.0	<0.0001
Treatment modality
Hemodialysis (%)	65.9	71.2	67.9	66.4	0.001
Cause of renal failure (%)	–	–	–	–	<0.0001
Uncertain	24.9	30.7	18.8	25.1	–
Diabetes	17.8	36.6	33.2	20.2	–
Renovascular disease	13.9	7.9	18.9	13.6	–
Other	14.0	8.1	10.8	13.4	–
Glomerulonephritis	11.6	8.1	11.5	11.3	–
Polycystic kidney disease	8.2	1.9	3.6	7.5	–
Chronic pyelonephritis	9.4	6.6	2.9	8.9	–
Area-level social deprivation (%)	–	–	–	–	<0.0001
quintile 1 (least deprived)	18.6	6.1	4.2	17.0	–
quintile 2	21.2	8.2	5.4	19.5	–
quintile 3	20.1	12.8	9.6	19.1	–
quintile 4	21.2	32.5	22.2	22.2	–
quintile 5 (most deprived)	18.8	40.4	58.5	22.2	–

SES and Attainment of Clinical Practice Guideline Standards

Results from both the unadjusted models and multivariable models sequentially controlling for the covariates are shown. The differences in attainment of various laboratory standards between SES categories are shown in Table 3.

Table 3.

Attainment of various standards by area-level social deprivation among White patients

	Logistic Regression models,a Odds ratio (95% CI)
	Step 1 unadjusted	Step 2 + patient factorsb	Step 3 + center effect
Hemoglobin ≥ 10 g/dl
Quintile 1	Reference	Reference	Reference
Quintile 2	0.88 (0.75 − 1.04)	0.92 (0.78 − 1.08)	0.92 (0.75 − 1.14)
Quintile 3	0.85 (0.72 − 1.00)	0.90 (0.76 − 1.06)	0.90 (0.78 − 1.05)
Quintile 4	0.79 (0.68 − 0.93)	0.86 (0.73 − 1.02)	0.88 (0.74 − 1.05)
Quintile 5	0.73 (0.62 − 0.86)	0.82 (0.69 − 0.97)	0.88 (0.72 − 1.08)
p value	0.003	0.2	0.5
p-trend	<0.0001	0.02	0.1
URR > 65%
Quintile 1	Reference	Reference	Reference
Quintile 2	1.00 (0.85 − 1.19)	1.02 (0.85 − 1.21)	1.03 (0.85 − 1.26)
Quintile 3	0.95 (0.81 − 1.13)	0.99 (0.83 − 1.17)	1.00 (0.87 − 1.15)
Quintile 4	0.90 (0.76 − 1.06)	0.94 (0.79 − 1.11)	0.95 (0.80 − 1.12)
Quintile 5	0.93 (0.78 − 1.09)	0.96 (0.81 − 1.14)	0.90 (0.75 − 1.09)
p valuec	0.6	0.8	0.7
p-trend	0.1	0.3	0.2
Phosphate < 5.6 mg/dl
Quintile 1	Reference	Reference	Reference
Quintile 2	1.02 (0.90 − 1.14)	1.05 (0.93 − 1.19)	1.07 (0.95 − 1.20)
Quintile 3	0.94 (0.83 − 1.06)	1.00 (0.89 − 1.14)	1.02 (0.91 − 1.14)
Quintile 4	0.87 (0.77 − 0.98)	0.97 (0.86 − 1.10)	0.96 (0.86 − 1.09)
Quintile 5	0.97 (0.86 − 1.10)	1.15 (1.01 − 1.30)	1.13 (1.01 − 1.28)
p valuec	0.07	0.06	0.03
p-trend	0.1	0.2	0.2
Calcium 8.8 to 10.4 mg/dl
Quintile 1	Reference	Reference	Reference
Quintile 2	0.97 (0.85 − 1.11)	0.98 (0.86 − 1.12)	1.01 (0.90 − 1.13)
Quintile 3	1.00 (0.88 − 1.14)	1.01 (0.88 − 1.16)	1.05 (0.91 − 1.21)
Quintile 4	0.99 (0.87 − 1.13)	1.00 (0.87 − 1.14)	1.05 (0.95 − 1.15)
Quintile 5	1.07 (0.93 − 1.22)	1.08 (0.94 − 1.24)	1.16 (1.06 − 1.26)
p valuec	0.7	0.7	0.02
p-trend	0.3	0.3	0.007
CaPo4 < 55 mg2/dl2
Quintile 1	Reference	Reference	Reference
Quintile 2	1.06 (0.94 − 1.20)	1.10 (0.97 − 1.24)	1.13 (0.98 − 1.28)
Quintile 3	1.01 (0.90 − 1.14)	1.08 (0.95 − 1.22)	1.09 (0.99 − 1.20)
Quintile 4	0.97 (0.86 − 1.09)	1.07 (0.95 − 1.21)	1.06 (0.95 − 1.18)
Quintile 5	1.07 ( 0.95 − 1.21)	1.25 (1.10 − 1.42)	1.25 (1.12 − 1.38)
p valuec	0.4	0.01	0.0004
p-trend	0.8	0.004	0.004
PTH < 300 pg/ml
Quintile 1	Reference	Reference	Reference
Quintile 2	1.07 (0.92 − 1.24)	1.07 (0.92 − 1.24)	1.04 (0.88 − 1.23)
Quintile 3	1.03 (0.89 − 1.20)	1.02 (0.88 − 1.19)	1.01 (0.89 − 1.25)
Quintile 4	1.05 (0.91 − 1.22)	1.04 (0.89 − 1.21)	1.03 (0.87 − 1.22)
Quintile 5	0.98 (0.84 − 1.14)	0.99 (0.84 − 1.16)	0.99 (0.87 − 1.14)
p valuec	0.8	0.8	0.9
p-trend	0.7	0.7	0.9

^aSeparate models were constructed for each of the standards. Covariates were sequentially added to the model from step 1 to step 3. Quintile 1 (least deprived); Quintile 5 (most deprived).

^bCovariates adjusted: age, sex, cause of end stage renal disease, year of start of RRT, RRT modality;

^cp values quoted are significance levels for heterogeneity.

RRT, renal replacement therapy; PTH, parathyroid hormone; URR, Urea reduction ratio; CaPo4, calcium phosphate product.

(1) Hb: SES was associated with poor attainment of the Hb standard in the unadjusted model only. Adjustment for patient factors attenuated the association although there was still a trend (p-value for trend = 0.02). Additional adjustment for centers further weakened the gradient, which was now consistent with chance.

(2) HD dose (URR): There was no association between SES and attainment of the URR standard in the unadjusted or adjusted models.

(3) Bone mineral disorder: There was no association between SES and attainment of the PO4 standard in the unadjusted model, but there was some evidence in the fully adjusted model that the patients from the most deprived quintile had better attainment of PO4 standards compared with those from more affluent areas, but there was no clear dose-response effect. The UK Renal Association guidelines have not set a lower limit for the PO4 standard. When we repeated the analysis using the K/DOQI targets for PO4 (3.5 to5.5 mg/dl), we did not find any association between SES and attainment of the PO4 standard (data not shown). There was no association between SES and attainment of Ca or CaPO4 standards in the unadjusted model, but there was some evidence in the fully adjusted model that patients from more deprived areas had better attainment of Ca and CaPO4 standards. There was no association between SES and attainment of the PTH standard in either the unadjusted or fully adjusted models.

(4) SES and center interaction: No significant interaction was observed between SES and center for any of the laboratory variables (data not shown).

Ethnicity and Attainment of Clinical Practice Guideline Standards

The ethnic differences in attainment of various laboratory standards are shown in Table 4.

Table 4.

Logistic regression models comparing attainment of various laboratory standards between ethnic groups

	Logistic Regression models,a Odds ratio (95% CI)
	Step 1 unadjusted	Step 2 + patient factorsb	Step 3 + area- level social deprivation	Step 4 + center effect
Hemoglobin ≥ 10 g/dl
White	Reference	Reference	Reference	Reference
South Asian	0.96 (0.81 − 1.14)	1.05 (0.88 − 1.24)	1.07 (0.90 − 1.28)	1.02 (0.84 − 1.23)
Black	0.62 (0.50 − 0.76)	0.66 (0.53 − 0.83)	0.68 (0.55 − 0.85)	0.57 (0.45 − 0.71)
p value	0.0001	0.001	0.002	<0.0001
URR > 65%
White	Reference	Reference	Reference	Reference
South Asian	1.36 (1.13 − 1.63)	1.57 (1.29 − 1.90)	1.57 (1.30 − 1.91)	1.78 (1.31 − 2.41)
Black	0.81 (0.63 − 1.03)	0.84 (0.65 − 1.09)	0.85 (0.65 − 1.09)	0.84 (0.62 − 1.13)
p value	0.0005	<0.0001	<0.0001	0.0009
Phosphate < 5.6 mg/dl
White	Reference	Reference	Reference	Reference
South Asian	1.20 (1.05 − 1.37)	1.38 (1.20 − 1.58)	1.35 (1.18 − 1.55)	1.26 (1.06 − 1.49)
Black	1.60 (1.31 − 1.94)	1.90 (1.55 −2.32)	1.83 (1.49 − 2.24)	1.62 (1.31 − 2.00)
p value	<0.0001	<0.0001	< 0.0001	<0.0001
Calcium 8.8–10.4 mg/dl
White	Reference	Reference	Reference	Reference
South Asian	0.77 (0.67 − 0.88)	0.79 (0.69 − 0.91)	0.78 (0.67 − 0.89)	0.77 (0.65 − 0.90)
Black	1.14 (0.92 − 1.40)	1.15 (0.93 − 1.42)	1.11 (0.90 − 1.38)	1.05 (0.82 − 1.35)
p value	0.0003	0.001	0.001	0.001
CaPo4 < 55 mg2/dl2
White	Reference	Reference	Reference	Reference
South Asian	1.27 (1.11 − 1.46)	1.44 (1.25 − 1.66)	1.39 (1.21 − 1.60)	1.26 (1.08 − 1.47)
Black	1.79 (1.46 −2.19)	2.11 (1.71 −2.60)	1.99 (1.61 − 2.45)	1.71 (1.40 − 2.09)
p value	<0.0001	<0.0001	0.001	<0.0001
PTH < 300 pg/ml
White	Reference	Reference	Reference	Reference
South Asian	0.58 (0.50 − 0.68)	0.59 (0.50 − 0.69)	0.60 (0.51 − 0.70)	0.62 (0.54 − 0.72)
Black	0.30 (0.24 − 0.36)	0.30 (0.25 − 0.38)	0.32 (0.25 − 0.39)	0.27 (0.22 − 0.33)
p value	<0.0001	<0.0001	<0.0001	<0.0001

^aSeparate models were constructed for each of the standards. Covariates were sequentially added to the model from step 1 to step 4.

^bCovariates adjusted: age, sex, cause of end stage renal disease, year of start of RRT, RRT modality.

RRT, renal replacement therapy; PTH, parathyroid hormone; URR, Urea reduction ratio; CaPo4, calcium phosphate product.

(1) Hb: Attainment of the Hb standard was lower only in Blacks compared with Whites in the unadjusted model, and this difference persisted even after adjustment for center and patient characteristics, including SES.

(2) HD dose (URR): In the fully adjusted model, the South Asians had better attainment of the URR standard than Whites. There was a trend toward lower attainment of the HD dose standard in Blacks, but this was consistent with chance variation.

(3) Bone mineral disorder: In the fully adjusted model, attainment of the PO4 standard was better among South Asians and Blacks compared with Whites. When we repeated the analysis using the K/DOQI targets for PO4 (3.5 to 5.5 mg/dl), the fully adjusted odds ratio for South Asians was 1.08 (95%CI 0.93, 1.24) and for Blacks was 1.27 (95%CI 1.09, 1.49). In the fully adjusted model, attainment of the Ca standard was lower for South Asians but no different for Blacks. In the fully adjusted model, attainment of CaPO4 standard was higher for South Asians and Blacks. Despite having better attainment of PO4 and CaPO4 standards, Blacks and South Asians had lower attainment of PTH standard compared with Whites.

(4) Ethnicity, SES and center interaction: No significant interaction was observed for any of the laboratory variables between ethnicity and SES, or between ethnicity and center (data not shown).

Discussion

This is the first multicenter study outside North America to describe ethnic and socioeconomic differences in attainment of clinical practice guideline standards for patients on dialysis. While this study found no evidence that patients from lower SES experienced inequity in attainment of standards after adjustment for patient and center characteristics, some crude differences were observed. For Hb, poorer area-level SES was associated with lower attainment of the Hb standard, but this may be mediated through patient and center factors, as after adjustment the trend was substantially reduced and nonsignificant. There are no studies that have reported similar associations for patients established on dialysis; however, our findings are contrary to a previous report of low hematocrit levels at the start of RRT among those who lacked private health insurance (3). Our study suggests that socioeconomic disparities in management of anemia may be reduced in a managed-care setting in a health care system funded by general taxation and free at point of care and/or we were able to adjust for other covariates to a greater degree. Moreover, there was some evidence that the most deprived patients had better attainment of Ca and CaPO4 standards, although the reasons for these findings are not clear.

This study also reports, however, significant ethnic differences in the attainment of various standards that persist after controlling for SES and center performance. These findings confirm earlier reports of ethnic differences in Hb and dialysis dose attainment (5–7) but take the analysis further by adjusting for SES and center and by providing, for the first time, information about attainment of bone mineral disorder standards according to ethnic group. Correction of anemia with EPO has been reported to improve survival (26,27) and quality of life (28,29) for patients on RRT. This raises the possibility that optimal correction of anemia among Blacks may further improve the survival (30,31) and quality of life advantage (32,33) reported among Blacks on dialysis. Due to insufficient data on EPO usage, it is not known if the lower Hb levels among Blacks were related to inadequate treatment or decreased responsiveness to EPO. A recent report (34) has suggested that African American nonsmokers exhibit diminished response to EPO compared with other races and, therefore, may need higher doses of EPO to achieve similar Hb levels.

The greater attainment of HD dose observed in this study among South Asians has been reported before by other researchers (7). This may be partly explained by better adherence to treatment as has been described in Japanese patients (35), or it maybe due to their smaller body size and urea distribution volume (given that HD dose [Kt/V] is inversely related to body mass [BMI]) (36). BMI data for our study population was incomplete, but South Asians in the UK and Europe have been reported to have lower BMI compared with Whites (37–40). In our study, Black race and area-level SES were not predictors of lower HD dose and, therefore, lower dialysis dose reported in Blacks in the US (5) may have been mediated through SES.

Blacks and South Asians had lower attainment of the PTH standard despite having better or comparable attainment of PO4 and CaPO4 targets. Previous studies have reported that Black dialysis patients have severe hyperparathyroidism (6,41) but did not include South Asians. In the general population, Blacks and South Asians have secondary hyperparathyroidism (42–44) due to skin pigmentation and reduced synthesis of 25-hydroxyl vitamin D₃ (44–45), and this may have been exaggerated with the development of renal failure. The UKRR does not collect data on use of vitamin D analogues or phosphate binders, or whether patients have had parathyroid surgery. However, there is indirect evidence to suggest that Blacks and Asians may have decreased end-organ response to calcitriol due to vitamin D receptor gene polymorphisms (46–47), and compared with Whites, Blacks have decreased skeletal sensitivity to PTH (48–49) and lower levels of fibroblast growth factor −23, a hormone that has been implicated in decreased synthesis of calcitriol and increased mortality on dialysis (50). The more severe hyperparathyroidism among Blacks and South Asians may therefore be a physiologic adaptive response to maintain bone turnover. Previous studies have reported associations between high PTH and all-cause mortality (51–52) and cardiovascular- and fracture-related hospitalizations (52), but this was observed only at PTH levels of >480 pg/ml (52), and >600 pg/ml (51), much higher than the KDOQI recommended target of 150 to 300 pg/ml. It is also not clear if the PTH levels at which the mortality, hospitalization, and fracture-risk increases vary between ethnic groups. Furthermore, in contrast to Caucasians, there is no correlation between bone turnover and PTH levels in African American patients, such that PTH levels are comparable between African Americans with adynamic bone disease and Caucasians with high bone turnover disease (53). Based on these findings, the KDOQI guidelines setting uniform target values for bone mineral metabolism for all races have been recently questioned (54–55).

This study found no evidence of any interaction between area-level SES and ethnicity similar to another study of mortality in peritoneal dialysis patients (56), but an interaction has been reported in hemodialysis patients in the US (13). On the other hand, the association between SES and health-related outcomes has been well reported to differ between ethnic groups in nonrenal chronic diseases (14–18). These studies were all community-based studies and it may be that socioeconomic factors play a differential role on health-related outcomes between ethnic groups only in the primary health care setting. Equally, it may be that such interactions between SES and ethnicity do not exist in renal diseases, and further studies on other outcomes in renal diseases are needed to confirm our findings.

There are limitations to this analysis. Area-level socioeconomic characteristics, when used as proxies for individual-level SES, could lead to misclassification of individual socioeconomic position. However, this misclassification is likely to be nondifferential and therefore should not alter the conclusions of the study. We included center variables to adjust for variations in clinical management, which may confound geographically patterned sociodemographic factors. However, it is possible that these models are “over-adjusted,” as centers may also act to some degree as proxies for area deprivation (if their catchment population is mostly socially deprived), and that this variable captures an additional deprivation effect in addition to our area-based measure. Laboratory variables were also missing for a small proportion of patients, but baseline characteristics of patients included and excluded from the analysis for each of the laboratory variable were comparable (data not shown). Contrary to the K/DOQI recommendations, the UK Renal Association standards have set a dichotomous cut off without including an upper or lower limit for certain standards such as Hb, Po4, and PTH. However, we used these standards in the context of clinical practice in the UK at the time of the study. Data on use of EPO, vitamin D analogues, and phosphate binders were either unavailable or insufficient, and parathyroid surgery is not captured by the UKRR. This information could have helped us better understand the reasons for differences observed between ethnic groups. It is possible that patients with suboptimal pre-ESRD care could have died earlier and the ESRD cohort included in this study may represent a survival bias with a regression to the mean of patients with more similar laboratory values linked to improved survival.

In conclusion, it is encouraging that area-level SES, after accounting for patient and center factors, is not an important determinant of attainment of clinical practice guideline standards in the UK. There was no consistent pattern of inequity in attainment of standards by ethnic group. Dialysis patients of South Asian descent experienced comparable or superior outcomes to White patients for most clinical practice guideline standards except Ca standard and hyperparathyroidism. Further studies are needed to examine if this could partly explain the superior survival reported previously in this ethnic group (30,31). Compared with Whites, Blacks had similar or superior outcomes for dialysis dose, Ca, and PO4 but lower attainment of Hb and PTH standards, differences that do not appear to be explained by area-level measures of social deprivation but may have a biologic explanation.

Disclosures

None.