Distribution of serum amyloid A and establishment of reference intervals in healthy adults

Abstract Background Serum amyloid A (SAA) plays a critical role in acute or chronic and is used in clinical laboratories as an indicator of inflammation. The elevated SAA is closely related to inflammation‐mediated diseases, such as liver diseases, autoimmune diseases, metabolism‐related diseases, amyloidosis, and tumors. However, there is no unified population reference interval for SAA. This study aimed to investigate the distribution of SAA in healthy Chinese adults 20‐79 years of age and to establish its population reference interval. Methods A total of 2365 healthy subjects met the requirements of this study. The levels of SAA were detected using an AU5821 automatic biochemical analyzer and its original reagents. According to the recommended methods of CLSI C28‐A3 and WS/T 402‐2012, the population reference interval of SAA was established using the unilateral 95th percentile (P95), and the 90% confidence interval of upper limits was calculated. Results The distributions of SAA levels were not significantly different between sexes (P> .05) and also did not differ by age (P> .05). Therefore, the population reference interval for SAA was established as an upper limit of 11.0 mg/L (90% confidence interval: 9.3‐12.3 mg/L) by using the method of latex immunoturbidimetry. Conclusions Serum amyloid A is closely related to the occurrence and progression of various diseases. The preliminary establishment of a population reference interval for SAA can fully exert its potential clinical value.


| INTRODUC TI ON
Serum amyloid A (SAA) plays n critical role in acute or chronic and is used in clinical laboratories as an indicator of inflammation. 1 Although both SAA and C-reactive protein (CRP) are acute-phase proteins, the detection of SAA is more conclusive than the detection of CRP in patients with viral infections, severe acute pancreatitis, and rejection reactions to kidney transplants. 2 The elevated SAA is closely related to inflammation-mediated diseases, such as liver diseases, autoimmune diseases, metabolism-related diseases, amyloidosis, and tumors. [3][4][5][6] Besides, in acute-phase reactions such as acute inflammation and trauma, the concentration of SAA in the blood can be rapidly increased by approximately 1000-fold within 5-6 hours under the stimulation of IL-1, IL-6, and TNF-α. 7,8 Therefore, SAA has important clinical value in the diagnoses, progression, and prognoses of diseases associated with inflammation.
The reference interval is an important indicator for judging whether the test results are normal. As early as the 1980s, the

International Federation of Clinical Chemistry and Laboratory
Medicine (IFCC) recommended the establishment of reference intervals for laboratories. 9 The China National Accreditation Service for Conformity Assessment (CNAS) CNAS:2012, 5.5.2 also clearly stated that the laboratory should specify biological reference intervals or clinically determined values, and even document the basis of this regulation. 10 However, due to the establishment of the reference interval, many factors such as the inclusion and exclusion criteria of the reference individual, large workload, high economic cost, and other factors are constrained and affected. 11 Therefore, the establishment and application of reference intervals for laboratory analytes are greatly limited. At present, the population reference interval of SAA is mainly derived from kit instructions. Because the reference interval is also susceptible to methodological differences, population characteristics and geographical differences, the reference interval provided by the reagent manufacturer has limitations, so a laboratory self-built reference interval is particularly important. 12,13 Unfortunately, so far there have been a few reports about the establishment of SAA reference intervals in humans. 14,15 Hence, the present study was conducted through recruitment, sampling, and measurement. Meanwhile, we collected data of healthy subjects, conducted preliminary analyses, and established the reference interval of SAA by using latex-enhanced immunoturbidimetry for healthy adults in China. These reference values can be used for the clinical application of SAA-related diseases.

| Study subjects
According to the principle of complete randomness, 2600 subjects who completed a physical examination at the Physical Examination

| Specimen collection
A peripheral venous blood sample (5 mL) was drawn in the morning.
The specimens were centrifuged at 1200 g for 10 minutes, and the detection of all analytes was completed within 4 hours. At the same time, 10 mL of fresh morning urine was collected, and the measurement of urine was completed within 2 hours. All instruments and equipment were maintained and calibrated according to the manufacturer's instructions, and the internal quality control was in control and external quality assessment was qualified prior to the detection of samples.

| Outlier test
According to the Dixon method recommended by the guidelines of CLSI C28-A3, 16 the detection results were sorted by size (from small to large) to calculate the range, R. The distribution was then used to calculate the difference (D) between the maximum value or the minimum value and its adjacent value. If the D/R was ≥ 1/3, then the maximum or minimum value was eliminated as an outlier; the remaining data were used in the above steps until all outliers were eliminated.

| Identification of biological reference interval groups
According to the recommended methods of CLSI C28-A3 and WS/T 402-2012 ("Establishment of reference interval for clinical laboratory test items"), 16 where and x 1 and x 2 represent the mean values from the two groups, s 1 and s 2 represent the standard deviations from the two groups, and n 1 and n 2 represent the sample sizes of the two groups. If Z is > Z*, then the difference between the reference intervals is statistically significant (P < .05), and the reference intervals for the two groups should remain separate. In contrast, if Z is <Z*, then the difference between the two reference intervals is not statistically significant (P> .05), and the intervals can be merged into one reference interval.

| Statistical analysis
The data were analyzed using SPSS Statistics, version 20 (IBM). The Kolmogorov-Smirnov test was used to analyze the normality of the data. Nonparametric data were expressed using the median and interquartile range (IQR), and the Kruskal-Wallis H test was used to compare across multiple groups. A nonparametric (ranked) method was used to establish reference intervals according to CLSI C28-A3.
The difference was statistically significant at P < .05.

| RE SULTS
A total of 2365 subjects were ultimately retained from the inclusion and exclusion criteria in this study, which included 1152 males and 1213 females. The clinical characteristics of the subjects were listed in Table 1. The Kolmogorov-Smirnov test showed a skewed distribution for the concentrations of SAA (P < .05; Figure 2). The distribution of SAA levels was not significantly different between males and females, which indicated that the concentrations of SAA were not associated with sex (Z = 0.30, Z* = 9.42, P> .05; Table 2). No significant age-related differences in the levels of SAA were found, which showed that the data of SAA were not correlated with age (P> .05; Table 3).
In accordance with nonparametric methods recommended by CLSI C28-A3 and WS/T 402-2012, and combined with the clinical value of SAA, the upper reference limit was very important, which was set at the 95th percentile. The population reference interval of the SAA was established as an upper limit of 11.0 mg/L (90% confidence interval: 9.3-12.3 mg/L) by using the method of latex immunoturbidimetry in this study.

| D ISCUSS I ON
Serum amyloid A is an important acute-phase reactive protein mainly produced by liver, which can eliminate pathogens in the body and facilitate disease recovery. 18,19 And thus, it is a very sensitive marker reflecting an acute inflammatory state. 20 This study combined the was mainly detected by immunoturbidimetry, enzyme-linked immunosorbent assay, or microsphere capture enzyme immunoassay [23][24][25] ; (b) there were some differences in population characteristics and regional characteristics, such as race, environment, and living habits. 26,27 All of the above parameters may have resulted in different results. The results of this study suggested that every laboratory should establish a SAA population reference interval that accounts for the regional population, factors in the regional environment, and detection system characteristics.
However, there were some limitations in this study. First, this study was a single-center study and all subjects were from coastal areas, so the population distribution was relatively single. Second, the age of the subjects in this study was not fully covered, and there were no data on the part of the subjects younger than 20 years old.
Thus, in future studies, we plan to adopt multicenter research and cover all age groups of the subjects to ensure that the research results are more representative.
In conclusion, SAA is a sensitive acute-phase reactive protein, which is closely related to the occurrence and progression of various diseases. Therefore, the preliminary establishment of a population reference interval for SAA can fully exert its potential clinical value.

ACK N OWLED G M ENTS
None.