Pseudohypernatremia and pseudohyponatremia: a linear correction.

BACKGROUND: Serum sodium is commonly measured by direct potentiometry (DNa), in blood gas panels, or indirect potentiometry (INa), in metabolic panels run on chemistry analyzers. Abnormal values of the serum non-water fraction interfere with INa, with low values causing pseudohypernatremia (INa > DNa) and high values causing pseudohyponatremia (INa < DNa). Previous attempts to derive a linear correction for the difference between INa and DNa (ΔNa) arising from non-water bias--using serum total protein (TP) or albumin (ALB) to represent the non-water fraction--have yielded inconsistent results, possibly owing to differences in sample inclusion criteria, analytic platforms and statistical approach. METHODS: We quantified the effects of TP and ALB on ΔNa in 774 critical care patients with closely timed metabolic and gas panels, adjusting for other known effects. RESULTS: ΔNa varied inversely with TP, ALB, and the glucose difference between chemistry and gas panels (ΔGlu), and directly with pH and bicarbonate. The effect of TP on ΔNa was essentially linear, but that of ALB was not; hence, further analysis focused on TP. By multiple linear regression, ΔNa decreased by 0.64 ± 0.06 mEq/L for each 1 g/dL increase in TP, adjusted for ΔGlu, pH, and regression to the mean; the TP effect was slightly steeper (0.69 ± 0.06 mEq/L), when adjusted for bicarbonate instead of pH. CONCLUSIONS: For each 1 g/dL rise or fall in TP, clinicians may find it useful to adjust INa by 0.7 mEq/L in the same direction in order to correct INa for non-water bias.

Pseudohypobicarbonatemia caused by an endogenous assay interferent: a new entity.

Serum total carbon dioxide, measured using a chemistry analyzer, and gas panel-derived plasma bicarbonate, calculated from the pH and partial pressure of carbon dioxide, often are used interchangeably for clinical purposes. When they disagree, there is a tendency to accept total carbon dioxide and discredit gas panel-derived plasma bicarbonate values. We report a patient who, during a 5-month hospitalization, had persistently low total carbon dioxide levels (12.4 ± 2.7 [standard deviation] mEq/L [12.4 ± 2.7 mmol/L]), measured using an enzymatic/photometric assay, and a high anion gap (19.2 ± 3.1 mEq/L [19.2 ± 3.1 mmol/L]), suggesting high-anion-gap metabolic acidosis, but who had gas panel-derived plasma bicarbonate (24.0 ± 0.9 mEq/L [24.0 ± 0.9 mmol/L]) and arterial pH values in the reference range. Organic anion levels in blood and urine were unremarkable. Negative interference with the enzymatic assay by the patient's serum was shown by the findings that total carbon dioxide level was 7.0 ± 0.1 mEq/L (7.0 ± 0.1 mmol/L) higher when measured using the electrode-based method than using the enzymatic method (P < 0.01), and the patient's serum, but not control serum, altered the reaction kinetics of the enzymatic assay by producing turbidity, resulting in an initial increase in absorbance and a falsely low total carbon dioxide value. The turbidity may have resulted from precipitation of 1 of 2 paraproteins in the patient's serum or an endogenous antibody binding with an animal protein included in the assay reagents. In summary, a discrepancy between total carbon dioxide level measured using an enzymatic assay and gas panel-derived plasma bicarbonate level was found to be the result of turbidity caused by an endogenous interferent with the total carbon dioxide assay, a novel artifact. When total carbon dioxide and gas panel-derived plasma bicarbonate values disagree, measurement error in total carbon dioxide level should be considered.

CKD screening and management in the Veterans Health Administration: the impact of system organization and an innovative electronic record.

At the beginning of this decade, Healthy People 2010 issued a series of objectives to "reduce the incidence, morbidity, mortality and health care costs of chronic kidney disease." A necessary feature of any program to reduce the burden of kidney disease in the US population must include mechanisms to screen populations at risk and institute early the aspects of management, such as control of blood pressure, management of diabetes, and, in patients with advanced chronic kidney disease (CKD), preparation for dialysis therapy and proper vascular access management, that can retard CKD progression and improve long-term outcome. The Department of Veterans Affairs and the Veterans Health Administration is a broad-based national health care system that is almost uniquely situated to address these issues and has developed a number of effective approaches using evidence-based clinical practice guidelines, performance measures, innovative use of a robust electronic medical record system, and system oversight during the past decade. In this report, we describe the application of this systems approach to the prevention of CKD in veterans through the treatment of risk factors, identification of CKD in veterans, and oversight of predialysis and dialysis care. The lessons learned and applicability to the private sector are discussed.

Graded interference with the direct potentiometric measurement of sodium by hemoglobin.

OBJECTIVES: Sodium concentration is measured by either indirect (INa) or direct potentiometry (DNa), on chemistry and gas panels, respectively. A spurious difference between these methods (ΔNa=INa-DNa) can be confusing to the clinician. For example, variation in serum total protein (TP) is well known to selectively interfere with INa. Red cells have been suggested to interfere with DNa, but both positive and negative interference have been reported. In this study, the effect of gas panel hemoglobin (Hb) on ΔNa was examined. METHODS: ΔNa was calculated in 772 pairs of closely-timed chemistry and gas panels (median: 4min. apart), retrospectively collected from our critical care units, with 1 pair per patient. Hb was treated as a categorical or continuous variable and tested for linear and non-linear effects, with adjustment for 3 known influences on ΔNa-TP, bicarbonate (tCO2), and the chemistry-gas panel glucose difference (ΔGlu). RESULTS: Hb ranged from 3.5 to 22.0g/dL [35-220g/L]. In categorical analysis, ΔNa increased with Hb, and the effect was essentially linear. By simple regression, ΔNa rose 0.06±0.03[SE]mmol/L per 1g/dL [10g/L] increase in Hb (p<0.05), but confounding was suspected because Hb also correlated (p<10-3) with TP, tCO2, and ΔGlu. Using multiple regression to adjust for the confounders, ΔNa rose 0.15±0.03mmol/L per 1g/dL [10g/L] rise in Hb (p<10-6). CONCLUSIONS: Increasing Hb spuriously decreases DNa and increases ΔNa. A linear correction for this artifact can reduce the discordance between INa and DNa, promoting their interchangeable use.

Effect of dialysis modality on plasma fibrinogen concentration: a meta-analysis.

BACKGROUND: Concentrations of plasma fibrinogen, a vascular risk factor, tend to be greater in patients on peritoneal dialysis (PD) than hemodialysis (HD) therapy, like concentrations of serum cholesterol, lipoprotein(a), and transthyretin, despite the substantial loss of protein during PD. Worse vascular outcome has been noted in PD patients compared with HD patients in several studies. METHODS: In this study, the mean difference in plasma fibrinogen levels (PD-HD) was quantified by means of meta-analysis of mean differences found in 12 cohorts with both PD and HD patients (set 1; N = 630) by using a fixed-effects model and meta-analysis of mean fibrinogen values reported in 30 cohorts of patients on a single dialysis modality (set 2; 8 PD cohorts, 22 HD cohorts; N = 2,096) by using a mixed model. RESULTS: On meta-analysis, the weighted mean difference (PD-HD) was 105 mg/dL (95% confidence interval [CI], 86 to 124 [3.1 micromol/L; 95% CI, 2.5 to 3.6]) in set 1 and 103 mg/dL (95% CI, 53 to 153 [3.0 micromol/L; 95% CI, 1.6 to 4.5) in set 2. CONCLUSION: Like other vascular risk factors, such as cholesterol and lipoprotein(a), plasma fibrinogen level is markedly greater in PD than HD patients, with an approximate difference of 100 mg/dL [2.9 mumol/L]. Different plasma reference ranges for fibrinogen need to be defined for PD and HD patients. The mechanism for the difference and the possible role of hyperfibrinogenemia in worsening vascular disease in PD patients deserve study.

Serum prealbumin is higher in peritoneal dialysis than in hemodialysis: a meta-analysis.

BACKGROUND: Although not widely appreciated, the reported concentration of serum prealbumin, like that of serum cholesterol, tends to be higher in patients on peritoneal dialysis (PD) than on hemodialysis (HD), despite the substantial loss of protein during PD. METHODS: The mean difference in serum prealbumin was quantified by meta-analysis of the mean differences found in six cohorts with both PD and HD patients (set 1; N = 639) using a fixed-effects model, and meta-analysis of the mean prealbumin values reported in 23 cohorts of unselected dialysis patients on a single modality (set 2; 9 PD cohorts, 14 HD cohorts; N = 12,256) using a mixed model. For comparison, the mean difference in serum albumin concentration between PD and HD also was estimated in sets 1 and 2 using the same methods. RESULTS: In set 1, the mean prealbumin difference (PD-HD) in the individual cohorts ranged from 3.6 to 14.7 mg/dL (P < 0.05 in five cohorts), and the weighted mean difference was 5.4 mg/dL (95% CI, 3.8 to 7.0 mg/dL). In set 2, weighted mean prealbumin was 8.1 mg/dL (95% CI, 5.2 to 10.9 mg/dL) higher in PD than in HD in the entire data set, and 6.9 mg/dL (95% CI, 5.2 to 8.6 mg/dL) higher in a sensitivity analysis that excluded two outlying HD studies. By contrast, weighted mean serum albumin concentration was significantly lower in PD than in HD in both sets 1 and 2; the mean difference was 0.25 g/dL (95% CI, 0.14 to 0.36 g/dL) in set 1 and 0.28 g/dL (95% CI, 0.14 to 0.42 g/dL) in set 2. CONCLUSIONS: Serum prealbumin level is approximately 6 mg/dL higher in PD than HD, perhaps due to the stimulation of hepatic synthesis by PD albumin loss, while serum albumin is approximately 0.3 g/dL lower in PD. Different reference ranges and clinical targets (such as, K/DOQI guidelines) are needed for PD and for HD.