Changes in osmol gap in chronic kidney disease: an exploratory study

Abstract

Aim: There is no data on osmol gap (OG) in chronic kidney disease (CKD) by stage and limited data on OG in adults on maintenance hemodialysis (HD). We aimed to examine the OG between different stages of CKD and to compare the OG pre- and post-HD in those on maintenance HD. Methods: We conducted a cross-sectional study of 67 patients. The participants were divided into six groups: Group 1—reference group (normal renal function), Group 2—CKD stage 2; Group 3—CKD stage 3; Group 4—CKD stage 4; Group 5—CKD stage 5 and not on dialysis. Group 6 were subjects on maintenance HD. Results: The means of OG ± standard deviation of Groups 1–6 were 15.25 ± 3.0, 20.73 ± 2.68, 22.85 ± 6.99, 24.11 ± 3.64, 25.15 ± 5.06, and 28.88 ± 3.45, respectively (p < 0.001). In the HD group, the difference between the pre-HD and post-HD OG was statistically significant (p < 0.001). Conclusion: There is a statistically significant upward trend for OG as CKD stage increases. The OG is elevated in patients on maintenance HD and is normalized by the HD. OG can be a valuable additional tool to suggest CKD stage and serve as a marker of dialysis adequacy.

• Chronic kidney failure
• end stage renal disease
• osmolality
• osmolarity
• renal dialysis

Introduction

Calculation of serum osmol gap (OG) can be of great value in the clinical diagnosis and management of certain forms of critical illness such as toxic alcohol ingestions.1 The OG is defined as the difference between the measured osmolality (the number of osmoles of solute per kilogram of solvent) and the calculated osmolarity (the number of solute per liter of solution).2

Only low-molecular weight ions and uncharged molecules that are present in relatively high concentrations contribute significantly to serum osmolality.1,2 OG, therefore, is an indication of unmeasured concentration of osmotically active particles, or solutes. Under normal circumstances, the OG is small as the osmolarity is a fairly good estimate of the osmolality, but in some conditions there are a significant number of abnormal substances present, which contribute to the total osmolality.1,2 Calculated osmolarity will be well below the actual measured osmolarity resulting in an elevated OG.2

Increased OG can be divided into two categories: increased OG with metabolic acidosis and without metabolic acidosis.2 Increased OG with metabolic acidosis can be caused by ingestions of ethylene glycol, methanol, formaldehyde or paraldehyde, diabetic or alcoholic ketoacidosis, lactic acidosis, and chronic kidney disease (CKD) without dialysis. Increased OG without metabolic acidosis can be due to ingestion of isopropanol, diethyl ether, mannitol, sorbitol or glycine infusion, and either severe hyperproteinemia or hyperlipidemia.2

There appears to be limited research of only three studies on the relationship of the OG and CKD. One adult study showed that OG is increased in patients with advanced CKD not on HD.3 Another study performed in children also found that OG is increased in children with advanced CKD not on HD.4 Also, OG is increased in children on maintenance HD when measured pre-dialysis which is subsequently then reduced by HD.4 A more recent adult study of those on maintenance HD found similar results of increased OG pre-HD that was reduced after HD treatment.5 However, there is a lack of data on OG in earlier stages of CKD as well as limited data of only one study5 on OG in adults on maintenance HD. This study has two objectives. First, we examine the OG between different stages of CKD. Second, we compare the OG pre- and post-HD in those on maintenance HD.

Methods

Patients and data collection

We conducted a cross-sectional study on 67 patients at a suburban New York City tertiary care center with a typical catchment area of low to low-middle socioeconomic status from January 2008 to June 2010. Those studied were aged 18 years and older, and were treated at the renal or general medical clinics at our institution or were on maintenance HD at our outpatient dialysis unit. The study was approved by the local institutional review board and all patients provided written informed consent prior to their inclusion in the study. The study was conducted following the ethical guidelines of the Declaration of Helsinki 2000 and Declaration of Istanbul 2008.

Serum osmolarity was calculated from the Dorwart and Chambers formula6 using the three major osmotic constituents of normal serum (sodium, urea, and glucose) which under ordinary circumstances contribute nearly all of the osmolality of the sample. In the absence of substantial increases in serum lipids or protein, mOsm/L and mOsm/kg of water may be used interchangeably.

(Abbreviations: Na—sodium, BUN—blood urea nitrogen)

This Dorwart & Chambers formula was found to yield most accurate results6 for OG calculation among a comparison of 13 different formulas, even though a simplified version (2[Na]) + [glucose]/18 + [BUN]/2.8) is widely used. The Dorwart & Chambers formula uses 1.86 instead of 2 as the coefficient for sodium. In experimental data, a coefficient for sodium is generally found to be somewhat less than two. This is often rounded off to 2 for the ease of calculation.6

Patients with different stages of CKD were screened. Glomerular filtration rate (GFR) was estimated with the use of the extended version of the Modification of Diet in Renal Disease (MDRD) equation7 that is based on serum creatinine, blood urea nitrogen (BUN), serum albumin, age, gender, and race: eGFR = 170 × (serum Cr)^−0.999× age^−0.176×(BUN)^−0.17× albumin^0.318× 0.762 [if female] × 1.18 [if African American].

Patients were excluded if they had any of the following: acute renal failure defined as more or equal to 25% decrease in baseline GFR, alcohol level >10 mg/dL (measured only if patient reported alcohol use in last 48 hours), serum glucose >300 mg/dL, triglyceride level >500 mg/dL, and total protein >8.4 g/dL measured simultaneously. Patients who were pregnant or who had a history or suspicion of toxic alcohol ingestion (ethylene glycol, isopropyl alcohol, formaldehyde, paraldehyde, diethyl ether, isopropanol, and methanol), history of prior suicidal attempt, illicit drug use or use of herbal remedies were also ineligible. Recent mannitol, and IV immunoglobulin use were assessed and patients were excluded if they received these infusions or had any urological procedures with irrigation performed within the past month.

Participants were divided into six groups. Group 1, the reference group, consisted of patients with normal renal function defined as having GFR ≥ 90 mL/min/1.73 m². Group 2 consisted of patients with CKD stage 2 (GFR 60 to 89 mL/min/1.73 m²). Group 3 consisted of patients with CKD stage 3 (GFR between 30--59 mL/min/1.73 m²). Group 4 consisted of patients with CKD stage 4 (GFR between 15 to 29 mL/min/1.73 m²). Group 5 consisted of patients with CKD stage 5 (GFR < 15 mL/min/1.73 m²) and not on dialysis. Group 6 were patients on maintenance HD. Group 6 was dialyzed with high-efficiency cellulose diacetate dialyzer for single use (Dicea, Baxter US, Deerfield, IL) three times a week. The OG in this sample was measured immediately pre- and post-HD treatment.

We collected demographic variables of age (years), sex (men/women), race/ethnicity (White, African American, Hispanic, or other) as well as diabetes status (no/yes). Serum electrolytes (mmol/L), BUN (mg/dL), creatinine (mg/dL), glucose (mg/dL), total protein (g/dL), albumin (g/dL), and serum osmolarity (mOsm/L) were analyzed on the same blood specimen. Sodium was measured by ion-selective electrode method on the ISE 1800 module (electrolyte measuring unit, Roche Diagnostics Corporation, Indianapolis, IN). Osmolality measurements were performed with an osmometer using the freezing point depression method (Micro Osmometer model, 3320, Advanced Instruments, Inc., Norwood, MA). For the HD group, a urea reduction ration (URR) was calculated as a measure of dialysis adequacy.

Statistical analysis

Descriptive statistics of mean and standard deviation (SD) were used to describe the continuous variables. Frequency and percentages were used to describe the categorical variables. Comparisons between the means of OG in all stages of CKD were done using the analysis of variance (ANOVA). Also, analysis of covariance (ANCOVA) was performed adjusting for BUN and glucose. Post-hoc Tukey comparisons compared the means of the CKD groups. The changes in pre- and post-HD OG was compared by a paired t test. All p values were two-sided and reported as being statistically significant on the basis of significance level of 0.05. Data were analyzed using SAS software, version 9.1 (SAS Institute, Cary, NC).

Results

There were 56 patients from our various groups who were compared to the 11 patients with normal renal function (reference group) (Table 1). Average age for all groups was 55.6 ± 15.1 years. Slightly less than half (46.3%, n = 31) were women. More than half (n = 53.7%, n = 36) were African Americans, only 16.4% (n = 11) were Whites. Slightly more than half (56.7%, n = 38) had diabetes. Total protein and triglycerides were not elevated in any of the groups. The HD group was adequately dialyzed as indicated by an URR of above 65%. None of the patients had dialysis disequilibrium syndrome (DDS).

Table 1. Baseline characteristics of the study participants.

	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
	(n = 11)	(n = 9)	(n = 17)	(n = 14)	(n = 7)	(n = 9)
	M (±SD) or % (frequency)	M (±SD) or % (frequency)	M (±SD) or % (frequency)	M (±SD) or % (frequency)	M (±SD) or % (frequency)	M (±SD) or % (frequency)
Mean Age (years) ± SD	47.2 ± 13.3	54.1 ± 18.9	63.8 ± 12.0	57.1 ± 12.7	52.4 ± 13.8	52.2 ± 17.8
Gender
Men	72.7% (8)	44.4% (4)	64.7% (11)	57.1% (8)	42.9% (3)	66.7% (6)
Women	27.3% (3)	55.6% (5)	35.3% (6)	42.9% (6)	57.1% (4)	33.3% (3)
Race
White (%)	0	11.1	23.5	14.3	14.3	22.2
AA (%)	45.5	55.6	47.1	64.3	28.6	77.8
Hispanic (%)	27.3	22.2	11.8	7.1	57.1	0
Other (%)	27.3	11.1	17.7	7.1	0	0
DM (%)	63.6	33.3	52.9	71.4	28.6	77.8
Glucose (mg/dL)	117.4 ± 30.5	121.3 ± 60.8	112.5 ± 37	114.8 ± 3.8	106.9 ± 32.2	141.2 ± 63.7
Sodium (mmol/L)	139 ± 1.7	137.7 ± 3	139.7 ± 2.8	139.6 ± 2.8	139.1 ± 2.5	137.1 ± 3
Total protein (g/dL)	7.5 ± 0.5	7.3 ± 0.4	7.4 ± 0.6	7.3 ± 0.7	6.3 ± 0.8	7 ± 0.6
Triglycerides (mg/dL)	121.9 ± 61.6	114 ± 55	139.5 ± 59.9	174.4 ± 108	163.1 ± 85.2	112.6 ± 69.9
BUN (mg/dL)	13.6 ± 6.8	18.1 ± 6.6	28.1 ± 6.8	49.6 ± 14.7	68.7 ± 17.6	63.2 ± 15.2
Creatinine (mg/dL)	0.76 ± 0.2	1.04 ± 0.3	1.74 ± 0.3	3.15 ± 1	4.81 ± 1	8.5 ± 2.8 (pre-HD)
GFR (ml/min/1.73 m^2)	126.2 ± 37.9	77.5 ± 10.5	43 ± 5.9	22.7 ± 3.8	12.0 ± 2.4	N/A
URR %	NA	NA	NA	NA	NA	73.7 ± 8

Notes: M—mean, SD—standard deviation, AA—African American, DM—diabetes mellitus, BUN—blood urea nitrogen, GFR—glomerular filtration rate as calculated by modification of diet in renal disease study group (MDRD) formula, URR—urea reduction ratio, NA—not applicable.

The means of OG (±SD) increase as the CKD stage is increasing with the lowest mean level for the normal OG reference group of 15.25 (±3.0) and the highest mean level of 28.88 (±3.45) for Group 6 (Table 2). These results were globally statistically significant at p < 0.001 (Figure 1). ANCOVA analyses adjusting for BUN and glucose also was globally statistically significant at p < 0.001. The post-hoc comparisons found that CKD stages 3, 4, 5, and 6 were significantly different from the reference group (p < 0.05). Also, stages 1, 2, and 3 were significantly different from stage 6 (p < 0.05). All other comparisons were not significantly different.

Figure 1. Mean osmol gap by groups (p < 0.001). Error bars indicate standard deviation values.

Table 2. Measured osmolality and calculated osmolarity by groups.

	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
	M (±SD)	M (±SD)	M (±SD)	M (±SD)	M (±SD)	M (±SD)
Osmolality measured (mOsm/kg)	294.2 ± 6.0	299 ± 9.5	308 ± 7.6	316.9 ± 6.9	323.4 ± 7.8	322.3 ± 8.9
Osmolarity calculated (mOsm/L)	278.9 ± 5.1	278.3 ± 8.6	285.2 ± 5.5	292.8 ± 7.4	298.3 ± 6.6	294.5 ± 8.7
OG (mOsm/kg)	15.25 (±3.0)	20.73 (±2.68)	22.85 (±6.99)	24.11 (±3.64)	25.15 (±5.06)	28.88 (±3.45)

Notes: M—mean, SD—standard deviation.

In the HD group, mean OG (±SD) pre- and post-HD were 28.88 (±3.45) and 14.76 (±4.54), respectively. The difference between the pre-HD and post-HD OG was statistically significant (p < 0.001) (Figure 2). There also was a statistically significant difference between the OG pre-HD in the HD group and the reference group (p < 0.001) (Figure 2). There was no statistically significant difference in the OG between the post-HD OG in the HD group and the reference group.

Figure 2. Mean osmol gap of reference, pre-HD and post-HD groups. Pre-HD and post-HD OG was statistically significant (p < 0.001). Pre-HD and the reference group was statistically significant (p < 0.001). No statistically significant difference between post-HD and reference group.

Discussion

We found a statistically significant upward trend for OG as CKD stage increased. We also found that the OG was the highest in the participants on maintenance HD, pre-HD and was statistically significant from the post-HD OG. However, the mean OG for the reference group and post-HD OG were similar and not statistically different.

We found that the mean of OG increases as the renal function worsens and the CKD stage increases. Other studies that looked at this relationship consisted of two adult studies3,5 and one pediatric study.4 These studies were composed of adult patients with advanced CKD3 (mean creatinine of 5.9 ± 0.9 mg/dL) and pediatric patients aged 2–16 years with mean BUN4 of 71.7 ± 38.9 mg/dL as well as both adult and pediatric patients on maintenance HD.4,5 OG was elevated in these patient populations.3–5 Our study adds to the literature in that we described and analyzed OG in different stages of CKD which to our knowledge has not been previously studied. The elevated OG in CKD likely results from retention of unidentified solutes.5

Our finding can be especially of value for patients with earlier stages of CKD when standard estimation equations for GFR (Cockcroft--Gault and MDRD) do not perform well.7 OG has been shown to be normal in patients with acute renal failure.3 However, as we show that OG increases with CKD stage, there is a potential usefulness of calculating OG to differentiate between acute and chronic renal failure or to identify acute on chronic renal failure, when a patient presents with elevated creatinine and an unknown creatinine baseline.

We also found that OG is elevated pre-HD and is decreased to a normal level of OG by HD. This is similar to both adult and pediatric studies of patients on maintenance HD4,5 which found that OG is increased in pre-HD and is significantly lowered in post-HD. However, the pediatric study4 had no reference group with normal renal function for comparison. Also, the adult study5 found that the mean OG post-HD was significantly higher than that in their control group. This differs from our finding that mean OG levels are not statistically different between post-HD OG level and the reference group. A reason for our lack of a difference in OG between our reference group and the post-HD group might be explained by our use of high-flux dialyzers (type of dialyzers used is not mentioned in the adult study5) and the different blood flows used during HD treatments (all our patients were dialyzed with blood flows of 380–400 mL/min as compared with the blood flows of 280–320 mL/min used in the adult study5) possibly providing better clearances.

Although we found that HD results in the rapid normalization of OG, this also may contribute to complications of HD. Although no one had DDS in our sample, DDS is a central nervous system disorder described in dialysis patients.8 It is characterized by neurological symptoms of varying severity that are thought to be primarily due to cerebral edema. The exact etiology of DDS is not well known, however it is thought to be due to rapid fluid shifts induced by urea removal9–11 known as the reverse osmotic shift hypothesis. Urea is generally considered to be an “ineffective” osmole, and therefore does not completely explain the etiology of DDS. In a subgroup of patients who develop DDS, this acute change in the OG might create an osmotic gradient that promotes water movement into the cells potentially resulting in cerebral edema and a variable degree of acute neurological dysfunction. We propose that the rapid change in an OG pre- and post-HD may play a role in the pathogenesis of the DDS.

We found a mean OG for our reference group of 15.25 mOsm/kg. The commonly accepted normal value for OG is less than 10 mOsm/kg. However, more recent studies show that the OG varies by laboratory setting based upon the formula and instruments used for calculation and measurement of components of OG.12,13 Khajuria et al. concluded that laboratory in each institution has to verify their results in a population of normal patients.13 In the study, we used a reference group to determine the normal OG.

This study has several limitations. First, the relatively small sample size may have resulted in certain post-hoc comparisons to be non-significant. Second, due to the relatively small sample size, we were not able to adjust in our analyses for the potentially relevant covariates. Third, this was done in a single medical center and may not generalize to other institutions. Fourth, we used the freezing point method for measuring osmolality. Our findings may not generalize to other institutions that do not use the freezing point method.

In conclusion, to our knowledge, this is the first study in adults to show that OG increases as the CKD stage increases and OG that is elevated in patients on maintenance HD is normalized by the HD. We suggest that OG, in addition to its known uses such as identifying certain toxic alcohol ingestions, has other clinical applications. This relatively simple calculation can be used as an additional tool to suggest CKD stage and in addition to the standard measures of KT/V and URR, can serve as a measure of dialysis clearance.

Declaration of interest

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