Basic information of participants
10 laboratories participated in the first EQA in 2021, and 28 laboratories from 13 cities in 6 provinces (Gansu, Shaanxi, Sichuan, Chongqing, Xinjiang, Yunnan) participated in the second EQA in 2022. The types of laboratories participating in the EQA included andrology laboratories and clinical laboratories, of which andrology laboratories accounted for more than 80%. More than 80% of the participating hospitals were tertiary hospitals. Currently, the main international SDF testing methods are TUNEL, comet test, SCSA and SCD, but all participants in this EQA used SCSA and SCD in China. SCSA method increasing from 50% in 2021 to 60.7% in 2022(Table 1).
Table 1
Basic information about laboratories participating in the external quality assessmen.
Basic information
|
2021
|
2022
|
Number of participants
|
10
|
28
|
Geographical distribution
|
|
|
Provinces
|
2
|
6
|
Cities
|
5
|
13
|
Type of laboratories
|
|
|
Andrology laboratory
|
8/10(80.0%)
|
25/28(89.3%)
|
Clinical laboratory
|
2/10(20.0%)
|
3/28(10.7%)
|
Hospital grade
|
|
|
Tertiary hospital
|
8/10(80.0%)
|
24/28(85.6%)
|
Secondary hospital
|
1/10(10.0%)
|
2/28(7.2%)
|
Primary hospital
|
1/10(10.0%)
|
2/28(7.2%)
|
Detection Method
|
|
|
SCSA
|
5/10(50.0%)
|
17/28(60.7%)
|
SCD
|
5/10(50.0%)
|
11/28(39.3%)
|
SCSA, sperm chromatin structure assay; SCD, sperm chromatin dispersion
The homogeneity and stability of samples
One-way ANOVA was used to evaluate the homogeneity of the samples. The P values of 2021A, 2021B, 2022A and 2022B were 0.102, 0.731, 0.492 and 0.537, respectively, all of which were greater than 0.05, indicating that the samples were homogeneous (Supplementary Table 1). The stability of samples was tested by independent sample T test. The P values of 2021A, 2021B, 2022A and 2022B were 0.757, 0.060, 0.332 and 0.556, respectively, which were all greater than 0.05, indicating that the results were stable when the samples were transported at -80℃ for 5 days (Supplementary Table 2).
The results of SDF
A large spread of results was obtained for the four samples, and the highest values were 13.7, 4.2, 8.0 and 4.0 times the lowest for 2021A, 2021B, 2022A and 2022B, respectively. The range of the high SDF value samples (2021B and 2022A) is larger than that of the low SDF value samples (2021A and 2022B)(Table 2).
The CVs were very high in samples with low SDF values, with 46.6% and 30.3% for 2021A and 2022B. The CVs of the samples with high SDF values were lower than those of the samples with low SDF values, 30.1% and 26.7% for 2021B and 2022A(Table 2).
Table 2
The mean, standard deviation (SD), range and Coefficient of variation (CV) of sperm DNA fragmentation results of four samples
Sample no.
|
Mean(%)
|
SD
|
Range(%)
|
CV(%)
|
2021A
|
13.1
|
6.1
|
1.6–21.9
|
46.6
|
2021B
|
46.8
|
14.1
|
14.7–60.4
|
30.1
|
2022A
|
39.7
|
10.6
|
6.5–51.7
|
26.7
|
2022B
|
23.1
|
7.0
|
9.4–37.8
|
30.3
|
Comparison of SCSA and SCD methods
The median results of the four samples tested by the SCSA were 12.0% (Q1-Q3:6.7–18.5%), 51.4% (Q1-Q3:33.0-59.2%), 43.6% (Q1-Q3:38.1–47.7%) and 25.0% (Q1-Q3:21.1–30.0%), respectively. The median results of the four samples detected by SCD were 15.5% (Q1-Q3:9.1–17.5%), 48.0% (Q1-Q3:37.7–54.6%), 40.0% (Q1-Q3:32.3–47.0%) and 23.0% (Q1-Q3:14.8–26.0%), respectively. Except for 2021A, the median results of SCSA were slightly higher than SCD for the remaining three samples (Fig. 2). Furthermore, there was no significant difference between the results of SCSA and SCD in 4 samples (p = 0.463, 0.465, 0.312, 0.145). The CVs were 58.4%, 39.5%, 23.3% and 30.5% for SCSA and 39.9%, 21.5%, 32.4% and 25.1% for SCD, respectively. Apart from 2022A, the CVs of SCSA were higher than those of SCD(Fig. 3).
The results of 10 participants in 2021 and 2022
The 10 laboratories that participated in the EQA in 2021 all participated in the EQA in 2022. In 2021, the CVs of low value sample and high value sample of 10 participants were 46.6% and 30.1%, and the CVs decreased to 32.5% and 22.7% in 2022.
The mean of bias of 2 samples from each of the 10 participants in 2021 was compared with the mean of bias in 2022, the mean of bias decreased for 8 participants, with the most significant one decreased from 80.5–6.0%. There was a slight but insignificant increase in the mean of bias in 1 participant. 1 participant had a significant increase in the mean of bias,, from 11.3–60.3%(Fig. 4). The mean bias of the 10 participants in 2021 was 24%, while decreasing to 16% in 2022.
Case analysis of participants with abnormal results
Case A: the flow cytometer voltage was set too low.
Case description
In 2021, the target values of the two samples were 13% and 55%, and one of the participants reported 2% and 15%, respectively.
Case analysis and solution
The results of 2 samples were low at the same time, which was considered as systematic error. The causes could be technician operating problems, reagent problems or instrument problems. After analyzing the cause together with the participant's technician, it was found that the boundary between the sperm population and the other population in the FSC/SSC dot plot was not clear, so the reason for the lower results was thought to be the low voltage setting of the FSC channel of the flow cytometer. If the voltage is set too low, the boundary of the sub-populations will be unclear (Fig. 5A), resulting in low detection results. In contrast, the boundary between sub-populations with normal voltage settings should be clear (Fig. 5B). Therefore, the participant's technician worked with the instrument's engineer to adjust the FSC channel voltage, and in 2022, the participant tested 2 specimens with DFI target values of 43% and 22%, and the results were 40% and 21%. Therefore, it is suggested that when using flow cytometry to detect sperm DFI based on SCSA method, first and foremost, the laboratory should set and verify the voltage, so that the different characteristic particle populations can be distinguished well.
Case B: The amount of samples added was excessive.
Case description
In 2022, the target values of the two samples were 43% and 22%, and one of the participants reported 19.6% and 6%, respectively.
Case analysis and solution
The results of both samples were simultaneously low and were considered to be a systematic error. When checking the technician's testing process, it was found that the amount of semen sample added by the technician was 25ul(Fig. 6A). If too much semen sample volume is added, the effect of acid denaturation of acid treatment solution will be weakened. Consequently, the technician adjusted the added sample volume to 5ul and then tested the sample again, and the result of 2202A increased to 41.5% (Fig. 6B). Therefore, it is suggested that when SDF is detected by flow cytometry, on one hand, the final concentration of SDF sperm should be regulated to 1–2 million per milliliter according to the WHO manual, and on the other hand, the amount of semen added should not be too much.
Case C: The results were obtained through a fixed gate of software.
Case description
In 2022, the target values of the two samples were 43% and 22%, and one of the participants reported 24% and 9%, respectively.
Case analysis and solution
The results of two samples declining at the same time. A search for reasons revealed that the participant used the fixed gate of the flow cytometry analysis software directly to obtain the results(Fig. 7A). Nevertheless, manual gating analysis of the raw data files of 2202A by the organizers yielded results of 44% (Fig. 7B). Although some flow cytometers contain automatic analysis software, manual verification of gating accuracy by laboratories is still recommended.
Case D: The flow rate was set too high.
Case description
In 2022, the target values of the two samples were 43% and 22%, and one of the participants reported 24% and 15%, respectively.
Case analysis and solution
The results are significantly lower for both samples, which is considered to be a systematic error. After investigation, it was found that the flow rate on the flow cytometer was set to high speed when testing the samples, which may cause two or more sperm to adhere to the sample tube at the same time, affecting the test results. The flow rate should be set to low speed(no more than 300 events/s)for SDF detection on the flow cytometer. When a single cell passes through, the FSC-H correlates with the FSC-A and cells appear on a diagonal in a FSC-H/FSC-A dot plot(Fig. 8A). However, when cells clump together, It shows an off-diagonal distribution(Fig. 8B).