Renal Function and Serum Albumin at the Start of Dialysis in 514 Chinese ESRD In-Patients

Abstract

Background. The dialysis population has grown rapidly in recent decades. Despite the high cost and poor outcomes of dialysis treatment for ESRD, there are scant data about the level of renal function and the relationship of renal function and serum albumin at the start of dialysis in Chinese ESRD patients. Method. We report the level of serum creatinine (Scr), glomerular filtration rate (GFR), and serum albumin (Salb) in 514 ESRD in-patients who began their dialysis treatment between January 2001 through December 2007 at two large dialysis centers in Changsha, Hunan, China. Data were obtained through reviewing the case records of all 514 patients. GFR was predicted by an equation developed from the Modification of Diet in Renal Disease Study. In addition, serum albumin was analyzed in relation to levels of predicted GFR. Results. The mean (SD) and median predialysis serum creatinine was 1121.92 ± 458.24 and 1032 μmol/L. The mean (SD) and median predicted GFR was 4.98 ± 2.24 and 4.47mL/min/1.73m². The proportion of patients with predicted GFR of >10, 5 to 10, and <5 mL/min/1.73m² was 3.7, 36.2, and 60.1%, respectively. The mean predicted GFR was significantly lower among younger patients, uninsured patients, unemployed or farmer patients, patients who were employed, students, patients who selected hemodialysis, patients with ESRD caused by diseases other than diabetes, patients with BUN above the mean, and patients with hemoglobulin beneath the mean. Compared with patients who started with GFR >5mL/min, the patients who started with GFR ≤5mL/min had significantly higher plasma urea and creatinine levels but significantly lower creatinine clearance (mL/min per 1.73m²) and parameters of nutritional status, such as serum albumin, body weight, and BMI. Conclusion. A wide variation existed in renal function at the initiation of dialysis in partial Chinese ESRD patients. Most patients start dialysis at very low levels of predicted GFR. The nutritional status in patients who start dialysis early was better than those in patients who start dialysis when GFR ≤ 5mL/min. Further studies are needed to analyze the impact of level of renal function and nutritional status at the start of dialysis on the outcomes of ESRD.

Keywords:

• serum creatinine
• glomerular filtration rate
• chronic renal failure
• serum albumin
• dialysis start

INTRODUCTION

The global population of ESRD patients has grown rapidly and is treated predominantly by dialysis in recent decades. The annual cost of treating patients on dialysis has also grown simultaneously. Despite the resources committed to the treatment of ESRD, these patients experience a significant morbidity and a reduced quality of life. Hospitalization rates for patients with ESRD are higher than those for age- and gender-matched comparative cohorts. In addition, the mortality among dialysis patients remains high. Despite the high cost and poor outcomes of dialysis treatment for ESRD, there are scant data about the level of renal function at the initiation of dialysis in Chinese ESRD patients. Studies showing an excess mortality in patients referred at a late stage point out that the level of renal function at start of dialysis should possibly be considered.[1–3] Renal function cannot be quantified by simple measurements of serum urea and creatinine, especially considering that creatinine clearance has been proved to greatly overestimate glomerular filtration rate (GFR) when renal failure is severe. Therefore, many nephrologists have suggested using GFR developed from the Modification of Diet in Renal Disease Study to estimate renal function. Currently, the decision to start dialysis remains empirical, and there are scant data regarding the level of renal function at the initiation of dialysis in Chinese ESRD patients. Moreover, a recent study showed that serum albumin concentration below 40 g/L at the initiation of dialysis was associated with decreased survival.[4] To prevent this, an earlier start of dialysis has been promoted. The objectives of the present study were to define the patterns of initiation dialysis and to analyze the relationship of different levels of GFR with parameters of nutritional status, such as serum albumin, body weight, and BMI.

SUBJECTS AND METHODS

Patients

The population for this analysis consisted of new 514 ESRD patients who initiated chronic dialysis from two large dialysis centers in Changsha, Hunan, P. R. China, between January 2001 and December 2007. To limit this analysis to patients whose first ESRD treatment modality was dialysis, patients who received a transplant or returned to dialysis were excluded. Patients who were younger than 15 years of age were also excluded.

METHODS

The investigation form was designed according to the HCFA 2728 form, which contains demographic data, insurance, comorbid condition, cause of ESRD, functional status, employment, treatment modality selected, and laboratory information. The laboratory values recorded should include serum creatinine, blood urea nitrogen (Bun), hemoglobulin, serum albumin, serum potassium, serum calcium, and serum phosphate. The obtained data should at least have contained age, sex, serum creatinine, BUN, serum albumin, or body weight, or was excluded as insufficient data. The MDRD study equation to predict GFR was adopted to calculate the GFR values of all 514 patients[5]:where GFR is in mL/min per 1.73 m² of body surface area, and serum (plasma) creatinine (Pcr) is in mg/L. Age is in years, SUN is in mg/dL, and serum albumin (Salb) is in g/dL.

Age was categorized in 15 year intervals for descriptive purpose and as a continuous variable for the multivariable models. Gender was considered to be dichotomous variable. Causes of ESRD were categorized as primary glomerular nephritis, diabetic nephropathy, hypertension, obstructive nephropathy, or other renal diseases. Insurance status was categorized as medical insurance, public fund-aided, partial public fund-aided, or self-afforded. Employment status was defined as presently employed, unemployed/farmer/student, or retired. Dialysis modality (hemodialysis or peritoneal dialysis) was defined as the anticipated long-term primary type of dialysis at ESRD start. Furthermore, the GFR values calculated by the MDRD formula were also categorized in 2 mL/min/1.73m² intervals for comparison purpose. Body mass index (BMI; measured in kg/m²) was estimated with the following formula:

Statistical analysis was performed using SPSS11.0 software. Univariate analyses for serum creatinine and for predicted GFR were performed using t-tests or analysis of variance (ANOVA) to compare means between levels of each variable. Correlation between predicted GFR and serum albumin was performed using pearson's correlation coefficient. Differences was considered significant at p < 0.05.

RESULTS

Demographic and Clinical Characteristics of the 514 Investigated Patients

Demographic and clinical characteristics of the 514 patients were shown in Table 1. Patients without any kind of insurance or medicaid accounted for 54.4% of the 514 patients, and the proportion showed a trend of decline from the year 2001 to 2007. Only 17.7% of the patients had medical insurance, whereas the proportion showed a trend of yearly increase during the same period. 45.1% of the patients were farmers or the unemployed. As for the dialysis modality, 91.8% of the patients selected hemodialysis, whereas only 8.2% selected peritoneal dialysis. Chronic glomerular nephritis was the largest single cause of ESRD, accounting for 59.7% of the 514 patients. Hypertensive nephrosclerosis was the second largest cause of ESRD, accounting for 10.9%, and diabetic nephropathy accounted for 6.8%. The proportion of patients with ESRD caused by glomerular nephritis showed a trend of yearly decline from the year 2001 to 2007, whereas that of patients with ESRD caused by hypertensive nephrosclerosis and diabetic nephropathy showed a trend of yearly increase during the same observing period. (see Table 2 and Figures 1 and 2).

Table 1 Demographic characteristics of the 514 investigated ESRD patients

Defined groups	Number of patients	Percentage (%)
All patients	514	100
Age (years)	514	100
15–30	88	17.1
30–45	171	33.3
45–60	159	30.2
60–75	91	17.7
>75	9	1.8
Gender	514	100
Male	289	56.2
Female	225	43.8
Body weight (kg)	514	100
10–59	286	55.6
59–70	170	33.2
70–83	48	9.3
>83	10	1.9

Table 2 Other characteristics of the 514 investigated ESRD patients

Defined groups	Number of patients	Percentage (%)
Economic resources	514	100
Medical insurance	91	17.7
Public fund	67	13.0
Partial public fund	71	13.8
Self-afforded	285	55.4
Employment	514	100
Presently employed	200	38.9
Unemployed/farmer	232	45.1
Student	15	2.9
Retired	67	13.0
Dialysis modality	514	100
Peritoneal dialysis	42	8.2
Hemodialysis	472	91.8
Causes of ESRD	514	100
Glomerular nephritis	307	59.7
Diabetic nephropathy	35	6.8
Hypertension	56	10.9
Obstructive nephropathy	45	8.8
Other renal diseases	71	13.8

Figure 1. Yearly trend of economic resources.

Figure 2. Yearly trend of the causes of ESRD.

Serum Creatinine and Predicted Glomerular Filtration Rate

The mean (SD) and median predialysis serum creatinine was 1121.92 ± 458.24 and 1032 μmol/L, respectively. The lowest (25%) and the highest (75%) quartiles were 793 and 1355 umol/L. The mean (SD) and median predicted GFR was 4.98 ± 2.24 and 4.47mL/min/1.73m², respectively. The lowest (25%) and the highest (75%) quartiles were 3.47 and 5.97 mL/min/1.73m², respectively. The distribution of predicted GFR is presented in Figure 3. The proportion of patients with predicted GFR of >10, 5 to 10, and <5 mL/min/1.73m² was 3.7, 36.2, and 60.1%, respectively. The mean serum creatinine was significantly higher and the mean predicted GFR significantly lower among younger patients, uninsured patients, unemployed, farmer patients, patients who were employed, students, patients who selected hemodialysis, patients with ESRD caused by diseases other than diabetes, patients with BUN above the mean, and patients with hemoglobulin beneath the mean (p < 0.001). The mean serum creatinine and predicted GFR was significantly higher in male than in female patients (p < 0.001 and p < 0.05, respectively; see Tables 3 and 4).

Figure 3. Distribution of predicted glomerular filtration rate in 514 investigated ESRD patients. Study population comprises Chinese ESRD patients who initiated dialysis between January 2001 and December 2007. The mean (SD) and median predicted GFR was 4.98 ± 2.24 and 4.47mL/min/1.73m².

Table 4 Mean serum creatinine and predicted GFR at start of dialysis in 514 investigated Chinese ESRD patients

	Patient number	Serum creatinine (μmol/L)	Predicted GFR (mL/min/1.73m^2)
All	514	1121.92 ± 458.24	4.98 ± 2.24
Financial aid
Medical insurance	91	932.09 ± 382.44	5.96 ± 2.62
Public fund	67	1047.05 ± 429.34	5.38 ± 2.32
Partial public fund	71	980.21 ± 353.01	5.31 ± 1.98
Self-afforded	285	1235.44 ± 478.09	4.49 ± 2.02
Employment
Employed	200	1097.63 ± 424.04	5.02 ± 1.98
Unemployed and farmer	232	1220.57 ± 466.40	4.44 ± 1.97
Student	15	1191.67 ± 429.88	5.43 ± 2.69
Retired	67	837.21 ± 412.32	6.63 ± 2.89
Dialysis modality
Peritoneal dialysis	42	805.26 ± 262.72	6.57 ± 2.57
Hemodialysis	472	1150.10 ± 461.47	4.84 ± 2.16
Cause of ESRD
Diabetes	35	751.98 ± 256.00	7.44 ± 2.73
Other causes	479	1151.58 ± 455.79	4.80 ± 2.10

*p < 0.001 (for all comparisons).

Table 3 Mean serum creatinine and predicted GFR at start of dialysis in 514 investigated Chinese ESRD patients

	Patient number	Serum creatinine (μmol/L)	Predicted GFR (mL/min/1.73m^2)
All	514	1121.92 ± 458.24	4.98 ± 2.24
Age (year)
15–30	88	1311.38 ± 512.72^†	4.65 ± 2.21^†
30–45	171	1202.84 ± 466.17	4.59 ± 1.85
45–60	155	1092.68 ± 387.89	4.80 ± 2.03
60–75	91	869.67 ± 372.33	6.16 ± 2.74
>75	9	785.94 ± 360.65	6.87 ± 2.89
Gender
Male	289	1200.15 ± 483.38^†	5.18 ± 2.29
Female	225	1021.44 ± 403.03	4.73 ± 2.16
Weight (kg)
10–59	286	1090.82 ± 417.22	4.84 ± 2.13
59–70	170	1216.12 ± 529.45	5.46 ± 2.19
70–83	48	1137.14 ± 380.50	5.15 ± 2.02
>83	10	988.98 ± 514.59	6.01 ± 2.33

*p < 0.05 (for all comparisons); ^†p < 0.001 (for all comparisons).

Serum Albumin and GFR

Patients were divided into two groups on the basis of their predicted GFR at the start of dialysis treatment (see Table 5). Three hundred and nine patients started dialysis treatment with a GFR ≤5mL/min. Compared with patients who started with a higher GFR, these patients had significantly higher plasma urea and creatinine levels. Creatinine clearance (mL/min per 1.73m²) was significantly lower; the parameters of nutritional status, such as serum albumin, body weight, and BMI, were also lower.

Table 5 Parameters of renal function and serum albumin: comparison of patients with GFR ≤5mL/min or >5mL/min

Parameter	GFR ≤5 mL/min per 1.73 m2)	GFR >5mL/min per 1.73 m2)
Number of patients	309	205
Age (yrs)	46 (23)	48 (22)
% diabetes	21 (6.79%)	14 (6.83%)
Plasma creatinine (mmol/L)	1204.65 ± 454.32	973.25 ± 348.09^†
Creatinine clearance (mL/min per 1.73 m2)	4.41 ± 2.21	5.34 ± 1.86
Serum albumin (g/L)	30.69 ± 4.22	38.53 ± 4.64
BW (kg)	55.32 ± 12.34	63.14 ± 11.25
BMI (kg/m2)	19.51 ± 2.92	22.67 ± 3.23

Values are presented as % or means (SD).

*p < 0.05, ^†p < 0.001 compared with GFR ≤ 5mL/min.

DISCUSSION

The present analysis of a representative sample of patients who were starting chronic dialysis treatment in a medium-sized city of China revealed that a wide variation in predicted GFR and serum creatinine existed at the initiation of dialysis, and a substantial fraction of the investigated patients started their dialysis with a very low level of renal function. Among 514 patients who began chronic dialysis between 2001 and 2007, with serum creatinine from 793 to 1355 umol/L, predicted GFR at start of dialysis ranged from 3.47 to 5.97 mL/min/1.73m². The average serum creatinine and predicted GFR were 1121.92 ± 458.24 μmol/L and 4.98 ± 2.24 mL/min/1.73m², respectively. This phenomenon in part reflects the widespread use of serum creatinine as a guide to the level of renal function in patients with chronic renal disease in China. In normal individuals, creatinine is excreted principally by glomerular filtration, with tubular secretion of creatinine accounting for only 5–10% of the excreted creatinine. Therefore, values for creatinine clearance closely approximate values for GFR. Urine creatinine excretion reflects muscle mass and, to a lesser extent, meat intake. In patients with chronic renal disease, renal tubular secretion of creatinine is enhanced; hence, creatinine clearance greatly overestimates GFR. On the other hand, urine creatinine excretion is reduced because of the restriction of meat intake, malnutrition, muscle wasting, and in addition, increased extrarenal elimination of creatinine resulting from degradation of creatinine by intestinal microorganisms. Therefore, using serum creatinine alone to estimate renal function in chronic renal disease may lead to a substantial overestimation of the level of GFR, especially in patients with reduced muscle mass.[6],[7] The MDRD equation takes into account the most important factors that affect the relationship between serum creatinine and GFR, namely, age, gender, race, and variables related to protein intake (BUN) and nutritional status (Salb). Routine use of the MDRD study equation to calculate predicted GFR would allow physicians to focus more clearly on the level of renal function in patients with chronic renal disease.

The patient population in our study appeared young, which might be related to the fact that the major cause of the ESRD in China is primary glomerular nephritis. The GFR value of younger patients is lower than that of other age groups, probably because they are stronger and have better endurance, and thus uremic symptom appears later.

The male patients had a significantly higher mean predicted GFR than that of the female patients. The reasons for these gender differences are not very clear. However, we found that women, as compared with men, were more likely to have low muscle mass and delay dialysis start because of being in worse position economically.

We also found an obvious association between GFR and insurance status, employment, and dialysis. Because of the unmatured medical insurance system, the proportion of self-afforded patients (the unemployed, farmers, and students) is relatively high, and many patients delay dialysis start because they lack the necessary economic resources. Thus, most patients chose hemodialysis as their first treatment modality simply for the purpose of relieving the uremic symptoms or other complications. This is the main cause of why hemodialysis patients and self-afforded patients had lower GFR values than peritoneal patients. On the other hand, patients who selected peritoneal dialysis were more likely to have received pre-ESRD care by a nephrologist and may have started dialysis in a planned and timely manner[8]; thus, their GFR value would be higher than that of hemodialysis patients. Finally, we also found significant associations between GFR and causes of ESRD. Glomerular nephritis is still the primary cause of ESRD in China, although its proportion had showed a yearly decline from January 2001 to December 2007. However, the proportion of hypertensive nephrosclerosis and diabetic nephropathy had showed a yearly increase during the same period. Patients with ESRD caused by diabetic nephropathy had a significantly higher mean predicted GFR than patients with ESRD because of other diseases. This is inconsistent with the widely held belief that diabetic patients require initiation of dialysis earlier because of the appearance of uremic symptoms at higher levels of GFR[9]; on the other hand, it may also reflect lower muscle mass in diabetic patients with chronic renal disease related to more severe malnutrition. Because the MDRD equation had not taken into account all patients with diabetic status, it thus could not reduce the difference between the GFR of diabetic patients and that of non-diabetic patients.

Compared to the data of the United States in 1999, great differences existed in the constituent ratio of the causes of ESRD; the primary cause of ESRD in 514 Chinese patients is chronic glomerular nephritis (59.7%), while in the United States, diabetic nephropathy accounted for the highest proportion (43%). This constituent ratio difference may relate to socio-economic development status, living standard, lifestyle, eating habits, and medical insurance system construction. The relative younger age of the 514 patients is probably related to the early onset of ESRD in glomerular nephritis. The mean predicted GFR in our study is lower than that reported in a United States study in 1999. The lower GFR among patients in local China could be explained by the great disparity in the construction of medical insurance system between China and United States.

Defining criteria on when to start dialysis to achieve the best patient outcomes is important and has been a neglected issue for a long time. The United States has changed its criteria in succession for several times.[10] The year 2002 DOQI guidelines recommended starting dialysis when GFR falls below 15mL/min, if there is uremia. However, the widely accepted opinion is that dialysis should be commenced before renal function falls below 10mL/min; this opinion had been supported by many studies.[11],[12] One important reason is that malnutrition is more likely to occur, when dialysis is commenced at GFR <10mL/min,[7],[11] and malnutrition usually leads to a higher rate of morbidity and mortality.[13] In our study, comparing patients who started dialysis with GFR of ≤5 mL/min per 1.73 m² with patients who started with higher levels of residual renal function showed significant difference in serum albumin, body weight, and BMI. This suggests that nutritional status was not similar in each of the patient groups. Serum albumin, body weight, and BMI were significantly lower in patients who started with GFR of ≤5 mL/min per 1.73 m². The results of our study suggest that the relationship between predicted GFR and nutritional status at the start of dialysis treatment may not be the same in different countries and populations. The IDEAL trial aims to determine whether it is better to commence dialysis with a creatinine clearance of 10–14 or 5–7 mL/min/1.73m², but it will not give an outcome till the year 2008.[14] When the aim of the timely start of renal replacement therapy is the prevention of malnutrition, clinical judgment combined with nutritional status may be more valuable than one based on GFR. Well-controlled prospective studies are necessary to determine the effects of serum albumin and GFR at the start of dialysis treatment on morbidity and mortality during renal replacement therapy.

In summary, in this study we showed that a wide variation existed in renal function at the initiation of dialysis in partial Chinese ESRD patients. Most patients start dialysis at very low levels of predicted GFR. Nutritional status in patients who start dialysis early was better than those in patients who start dialysis when GFR ≤5mL/min. Taken together, the data suggest taking the unique GFR prediction formula in accordance with the body build of the Chinese population, if possible, to evaluate renal function. We recommend an earlier start of dialysis in patients with ESRD and establishing a widely covered medical insurance system to avoid the delay of dialysis start due to the lack of necessary economic resources.

DECLARATION OF INTEREST

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Fuyou Liu and Hong Liu contributed equally to this work.