The effect of storage time at −20°C on markers used for assessment of renal damage in children: albumin, γ-glutamyl transpeptidase, N-acetyl-β-D-glucosaminidase, and alpha1-microglobulin

Abstract

Objective

The objective of this study is to examine the influence of storage time at −20°C on the concentration of albumin, γ-GT, NAG, A1M, and creatinine in a large sample of healthy children.

Material and Methods

The New England Children’s Amalgam Trial followed 534 children – aged 6–10 at baseline - for 5 years, with annual urine collections. Urine samples were analyzed for creatinine, albumin, γ -GT, NAG, and A1M concentrations. Repeated measures ANCOVA was used to model the effect of storage time on these concentrations.

Results

γ-GT concentration decreased significantly with storage time at −20°C. There was also a limited decrease in NAG. Albumin, A1M, and creatinine concentrations did not appear to be affected by storage time at −20°C.

Conclusions

If it is necessary to interpret results from samples stored for a long time at − 20°C, it is advisable to account for storage time in statistical models.

Introduction

The excretion of various proteins in urine is widely used for assessment of kidney function, in clinical practice as well as in epidemiological studies of populations exposed to nehprotoxic compounds, for example mercury (1, 2). Albumin is the commonly used marker of glomerular integrity, while low molecular weight proteins, or enzymes, such as beta-2- microglobulin, alpha1-microglobulin (A1M; also called protein HC), N-acetyl-β-D-glucosaminidase (NAG), or γ-glutamyl transpeptidase (γ-GT), are used for assessment of effects on renal tubular cells.

In many cases, especially in epidemiological studies, it is necessary to store the urine samples for long periods of time before analysis. Storage at −20°C is common, although many recent studies have found detrimental effects at this temperature and recommend either storage at 4°C with immediate analysis or storage at −70°C long term.

We found fifteen studies of albumin yielding varying results. Nine studies (3–10) found a decrease after storage at −20°C for as little as one week (though sometimes not until after a longer time). Furthermore, two studies (11, 12) found a decrease only for samples with low/normal albumin excretion, but not for those with higher (abnormal) excretion. On the other hand, four studies (13–17) found no decrease after 2 months, 6 months, 2 years, and 2 years, respectively. However, these studies were all done on diabetics or patients with renal disease (who would have abnormally high albumin excretion). Therefore, these results do not necessarily contradict the studies performed on healthy subjects, which find an effect of storage.

Two studies (18, 19) have found a decrease in γ-GT concentration with storage at − 20°C. The former found a 60% drop in γ-GT concentration after just one day of storage. They found γ-GT to be best preserved when urine samples were stored at −70°C after initial centrifugation.

Two studies (8, 9) of NAG showed a significant decrease after 6 months, and one study (19) concluded that approximately 15% of NAG concentration is lost in the process of freezing/thawing, but there are no further decreases over time. On the other hand, another study (20, 21) of NAG found no effect of storage at −20°C for one month, but found a 50% decrease after one year.

Two studies of A1M found a significant effect of storage at −20°C. Klasen (15) found a decrease in concentration over two years of storage. Tencer (16, 17) found a non-significant drop of 14% after six months, and a significant decrease of 22% after one year.

Finally, Schultz (9) found that creatinine concentration was unaffected after 6–8 months of storage at −20°C, and Manley (8) found that creatinine concentration was relatively stable after 6 months storage, but significantly decreased after 2 years.

Overall, most studies find that the concentrations of these kidney function markers are underestimated after storage at −20°C. All such studies have been performed on adults, with the exception of one study of albumin (12) and one study of albumin and NAG (9) on diabetic children. No such studies have been conducted on healthy children.

The aim of the current analysis was to examine the influence of storage time at −20°C on the concentration of albumin, γ-GT, NAG, and A1M, as well as creatinine, in a large sample of healthy children.

Materials and Methods

These data were available as part of the New England Children’s Amalgam Trial (NECAT) (22, 23). The NECAT study was designed to examine effects of amalgam dental fillings in 534 children in Boston and Maine, aged 6–10 at the beginning of the study, for 5 years. Outcome measures of this trial included effects on the kidney, as measured by creatinine-corrected albumin, γ-GT, NAG, and A1M. Eligibility criteria for the trial included no evidence of kidney disorders, and the sample was gender balanced and racially diverse (see Table 1). The study was approved by the institutional review boards of all participating sites, and it conforms to the provisions of the Declaration of Helsinki. Written parental consent and child assent was obtained for all participants.

Table 1.

Demographic Information on the 534 children

		Mean (SD) or N (%)
Age at beginning of 5-year study		7.9 (1.4)
Sex	Male	247 (46.3%)
	Female	287 (53.8%)
Race	Non-Hispanic White	323 (62.1%)
	Non-Hispanic Black	98 (18.9%)
	Hispanic	38 (7.3%)
	Other	61 (11.9%)

The original protocol called for collection of timed overnight samples, with spot daytime samples collected for those children who did not provide an overnight collection. However, as the children got older, compliance with the overnight collection become more problematic and a decision was made to switch to spot samples at the clinic for all children. Therefore, the proportion of samples collected overnight vs. daytime varied over the course of the study, from 92% overnight vs. 8% daytime at baseline to 4% overnight vs. 96% daytime at year 5.

γ-GT and creatinine concentrations were measured yearly for all NECAT children (53% overnight vs. 47% daytime samples overall); albumin, NAG, and A1M concentrations were measured at years 3 and 5, with additional measurements for 57 children in a substudy at year 4 (12% overnight and 88% daytime samples overall).

Urine specimens were sent to Rochester General Hospital and Strong Hospital clinical laboratories in Rochester, NY for analysis of creatinine and γ-GT, as well as to the Sahlgrenska University Hospital in Goteborg, Sweden for analysis of creatinine, albumin, NAG, and A1M. When the specimen was not sufficient for both, priority was given to the laboratories in Rochester, NY. Creatinine concentration was measured in both places for use in the creatinine-adjustment of the concentrations of the kidney markers; however, the creatinine concentration reported here is that from the laboratories in Rochester, NY. Creatinine was determined by the photometric ‘Jaffe’ method.

Specimen storage

Urine samples were stored frozen at −20°C until analysis. The median storage time for the laboratories in Rochester, NY was 71 days, with only 7% of samples stored for longer than 6 months. The median storage time for the laboratory in Goteborg, Sweden was 6 months for albumin and NAG and 9 months for A1M. Nine percent of the samples were stored for longer than one year before analysis of albumin and NAG, with 32% stored longer than one year before analysis of A1M.

Assay methods

γ-glutamyl transpeptidase (γ-GT) was measured using the DADE BEHRING Dimension clinical chemistry system: Flex reagent cartridge.

N-acetyl-β-D-glucosaminidase (NAG) was determined with an automated photometric method based on the formation of ‘3-cresol purple’ at the reaction catalysed by NAG, using reagents and calibrator from Roche Diagnostics (detection limit 0.1 U/L).

Urinary albumin and alpha-1-microglobulin (A1M) concentrations were determined by automated nephelometric immunochemical methods using reagents and calibrator from Beckman Coulter. Additional internal reference samples were used in each analytical run. The detection limits were 2.4 mg/L for albumin and 4 mg/L for A1M. However, since several previous studies on A1M have been performed using antibodies from Dako (24), and the concentrations in these two commercial antigens (’calibrators’) differ, we corrected our A1M concentrations. Using the Beckman Coulter antigens results in 40 % higher levels compared to those obtained using the Dako reagents. In order to be able to compare our data with some previous studies, we divided our A1M levels in mg/L by 1.4.

Special measures were taken to prevent inhomogeneous albumin concentrations in the thawed urine samples. The frozen samples were thawed in a water bath (37°C), mixed and kept in the water bath for another 30 minutes. After mixing again, 10 µL of Tween20 was added to 1 mL of urine. Samples were then kept cool overnight, gently mixed on the following day, then placed in the 37°C water bath for 3 × 60 minutes, and gently mixed between every 60 minutes period. After this procedure they were analysed within 3 days.

Statistical analysis

Repeated measures analysis of covariance models, with compound symmetric variance structures, were fit to determine the effect of storage time on the concentrations (mg or U/L) of creatinine, albumin, γ-GT, NAG, and A1M. Log transformations were used when examination of the model residuals indicated that it was needed. The model for albumin excluded samples with microalbuminuria (albumin > 30 mg/g creatinine, 12% of samples), since these samples would have too large an effect on the results. All models controlled for age, sex, race, and lean body mass. Lean body mass was calculated as weight*(1 - % body fat), with body fat measured by a body fat scale (model TBF-551). We also controlled for urinary collection time (overnight vs. daytime sample) and creatinine concentration (except for creatinine models) (25).

To verify overall findings, sub-analyses were performed excluding storage times greater than one year. Additionally, the effect of storage time on the detectability of albumin, NAG, and A1M concentrations was examined by logistic regression.

Results

Table 1 shows background demographic data on the 534 children, and Table 2 shows descriptive data on the kidney marker concentrations, their corresponding creatinine-corrected levels, and storage times. Creatinine and γ-GT were almost always detectable, while levels of A1M were mostly below the limit of detection (27% detectable). Albumin and NAG were 88% and 93% detectable, respectively.

Table 2.

Concentrations and excretions for the kidney markers and their median storage time^a

	N	Median (Range)	Median Storage Time in Days (Range)
Creatinine
Concentration (g/L)	2909	1.2 (0.1 – 4.2)	71 (4 – 1007)
Albuminb
Concentration (mg/L)	813	7.2 (ND – 1320.0)	177 (42 – 468)
Excretion (mg/gC)	718	7.4 (1.5 – 773.1)
γ-GT
Concentration (U/L)	2912	26.0 (0.0 – 224.0)	71 (4 – 1007)
Excretion (U/gC)	2909	26.7 (0.0 – 246.7)
NAGb
Concentration (U/L)	813	1.2 (ND – 9.2)	176 (22 – 468)
Excretion (U/gC)	754	1.4 (0.1 – 7.8)
A1Mb
Concentration (mg/L)	813	ND (ND – 17.0)	262 (42 – 563)
Excretion (mg/gC)	223	4.7 (1.1 – 29.5)

γ-GT = γ-glutamyl transpeptidase, NAG=N-acetyl-β-D-glucosaminidase, A1M= alpha-1-microglobulin, ND=not detectable.

^aCreatinine-corrected excretions exclude samples with non-detectable concentrations.

^bAlbumin, NAG, and A1M were not measured in all years.

Figure 1 shows plots of (detectable) creatinine, albumin, γ-GT, NAG, and A1M concentrations vs. storage times. Table 3 reports the results of repeated measures analysis of covariance models of the detectable kidney function measures on storage time. Creatinine concentration remains unchanged with storage at −20°C. γ-GT concentration decreases significantly with storage time. NAG concentration also decreases significantly with storage time; however, this decrease is no longer significant when samples with storage times greater than one year are excluded (though the slope remains similar). Samples stored for longer are significantly more likely to have non-detectable NAG (p<0.001). No significant effect of storage time on albumin or A1M concentrations was found.

Figure 1.

Concentrations vs. Storage Time. (A) Creatinine, (B) Albumin (excludes cases with microalbuminuria), (C) GGT = γ-glutamyl transpeptidase, (D) NAG=N-acetyl-β-D-glucosaminidase, and (E) A1M= alpha-1-microglobulin. Regression line is log-transformed for GGT.

Table 3.

Repeated measures analysis of covariance models of kidney markers on storage time, controlling for age, sex, race, lean body mass, creatinine concentration (except for creatinine), and sample time (overnight vs. daytime sample).

	Storage Time (days)
	Coefficient (SE)	p-value
Creatinine
Concentration (g/L)	0.00015 (0.00015)	0.30
Albumina
Concentration (mg/L)	−0.00038 (0.0037)	0.92
γ-GT
Concentration (U/L)b	−0.0028 (0.00022)	<0.001*
NAG
Concentration (U/L)	−0.0010 (0.00044)	0.02c*
A1M
Concentration (mg/L)	−0.00043 (0.0011)	0.71

γ-GT = γ-glutamyl transpeptidase, NAG=N-acetyl-β-D-glucosaminidase, A1M= alpha-1-microglobulin, SE=standard error.

^*statistically significant.

^aexcludes cases with microalbuminuria (12%).

^blog-transformed.

^cin a re-analysis excluding cases stored for over one year, p=0.12 (coefficient (SE) = −0.00090 (0.00057).

Based on model results, in a daytime sample from an average child (age 12), we can expect a 40% decrease in γ-GT concentration after 6 months of storage at −20°C, with a 64% decrease after 1 year. For NAG concentration, we estimate a 9% decrease after 6 months and an 18% decrease after 1 year.

Discussion

We have shown that storage time at −20°C has a large effect on γ-GT concentration, and, to a lesser extent, NAG concentration. Albumin and A1M do not appear to be affected by storage time at −20°C, apart from a possible decrease by the freezing/thawing procedure per se, irrespective of storage time. Additionally, we found no effect of storage on creatinine concentration, which suggests that both the concentrations and creatinine-adjusted levels of renal markers will be affected to a similar extent. Analyses of our creatinine-adjusted levels do in fact confirm this (data not shown). We found little information in the literature on the effect of long term storage at −70°C on these biomarkers.

Although other studies have found that albumin and A1M decrease with storage at −20°C, we do not find a significant decrease with time of storage. One possible reason for this is that the process of freezing and thawing the samples make them inhomogeneous, decreasing the recovery of these proteins in the analysis, while there is no substantial further breakdown with time. In the present study special measures were taken to dissolve and recover all proteins after thawing. Another possibility is that because our study did not analyze the same samples over time, we may lack power to detect a significant decrease with time. Inconsistent results in the literature regarding albumin may be due to assay-dependent differences in reactivity to fragments of albumin (26).

Although we were not able to analyze the same urine samples at various storage times, our conclusions are relevant to large epidemiological studies that store urine samples for varying amounts of time before analysis. Additionally, our results extend the conclusions of previous research on adults to healthy children.

Based on the conclusions of our study and the studies of others, we recommend storing urine samples at −70°C, if storage is necessary (20). We also recommend future study of storage at −70°C to show the need for standardized freezing procedures of no less than −70°C. Furthermore, if it is necessary to interpret results from samples stored at −20°C, it is advisable to account for storage time in statistical models.