Background
The People’s Republic of China enacted a new law on the prevention and control of infectious diseases in 2004 [
1]. Under the terms of this law, malaria is categorized as a B Class infectious disease, and every administrative county in China is responsible for the prevention and control of malaria and for reporting malaria epidemic information. In addition, epidemic information reporting should be procedural and standardized with patterns and time limits. The China Information System Disease Control and Prevention (CISDCP) addressed the need for efficient collection and transmission of epidemic information on infectious diseases in China from county-level to prefecture-level, province-level, and national-level institutions of disease prevention and control [
2]. The 24-h requirement to report the initial diagnosis of a malaria case in CISDCP is especially beneficial in advancing the timely collection of data to accurately analyse the malaria epidemic situation in China [
3‐
5]. The number of malaria cases in China has been reported and published from CISDCP system in the past decade [
6‐
8].
However, effective collection and analysis of malaria epidemic data cannot be separated from the accurate diagnosis of each malaria case [
9]. External quality evaluations [
10] or establishment of regional reference laboratories [
11] are two practical methods to improve the quality of malaria laboratory diagnoses. Ethiopia conducted a month-long tracking and evaluation of
Plasmodium microscopy skills on 60 laboratory personnel in 2015, and only 8.3% (5/60) of the participants reached the professional level required to correctly identify
Plasmodium species [
9], and the agreement rate of
Plasmodium infection was only 71.2% (198/276) [
12]. However, the accuracy rate of species identification was significantly improved after personnel received training in
Plasmodium microscopy skills (OR = 7.0) [
12]. As a non-malaria-endemic country with an average annual diagnosis of 1500 imported malaria cases [
13‐
17], the United States not only pay attention to timely reporting of malaria diagnosis results, but also ensures diagnostic accuracy and proper identification of
Plasmodium species through cross-checking between laboratories [
18]. Only 79.1% (706/893) of the 893 malaria cases initially diagnosed in Belgium were confirmed to be the correct species by microscopic diagnosis in the national-level reference laboratory between 2013 and 2017 [
11]. Therefore, the World Health Organization (WHO) advocates the use of cross-examination in reference laboratories to improve the diagnostic quality of laboratories in malaria endemic areas [
19].
The Yunnan Province Malaria Diagnostic Reference Laboratory (YPMDRL) was formally certified as a member of the China Malaria Diagnosis Reference Laboratory Network in 2012 [
20], and began to assume the responsibility of re-testing of malaria cases initially diagnosed [
21] and assessment of negative samples from county-level laboratories in Yunnan Province, which it was confirmed that there was no false negative diagnosis from 2016 to 2018 (Additional file
1). It has been operating normally for nearly eight years and has re-tested the most malaria cases as well as the last three indigenous malaria cases during the stage of Elimination Malaria Programme (EMP) in China [
6‐
8]. The re-testing results of previous years have also been used to revise the
Plasmodium species initially diagnosed by the county-level laboratories in Yunnan Province. In order to introduce the re-testing work of YPMDRL to the peer and explore ways to ensure the continuous stability and improvement of laboratory diagnostic capacities in Yunnan Province after EMP, this paper conducted this systematic analysis to verify the accuracy of laboratory diagnostics against re-testing in Yunnan Province from 2013 to 2018.
Methods
Ethics statement
The study was approved by the Yunnan Institute of Parasitic Diseases and Ethical Committee. Genetic testing was performed on stored blood samples obtained as part of the routine diagnostic work-up of patients with fever who were suspected to have malaria. Although there was no risk and the data processing after sample collection was done anonymously, informed consent was obtained.
Data collection of initial malaria diagnoses
Data was retrieved from the CISDCP database through the infectious disease report management module. From 2013 to 2018, the “report place” and “entry time” as search terms were used to search for cases of malaria disease on Monday of each week. Exported data included the case ID number, diagnosis time, disease type, initial diagnosis institution, and case report place. According to the epidemiological investigation, the sources of malaria infection were determined as follows: indigenous infection cases included those who had no history of travel to epidemic areas outside Yunnan Province within 30 days before the onset of malaria; imported infection cases included those who had a history of migration from malaria endemic areas, such as Myanmar and Africa, within 30 days of malaria onset.
Blood sample collection and transport
Two peripheral blood samples from every malaria patient with a clinical attack were collected by county-level malaria diagnostic laboratories in their administrative jurisdiction. The blood samples were used for microscopic examination in the county-level laboratories to initially diagnose for malaria cases, and were used to prepare another’s blood smears and dried blood spot samples for re-testing in YPMDRL. The latter were sent twice each month to YPMDRL by postal service. When there are few cases, the blood samples may be sent at any time to YPMDRL.
YPMDRL re-testing malaria cases samples
Blood smears from the initially diagnosed malaria cases were examined by microscopy at the YPMDRL according to the methods of Dong et al. [
21] and Zhu et al
. [
22], and others [
23], and the four species of human
Plasmodium were identified according to their morphological differences, as seen in thin blood smears. According to the methods of the literature [
21,
23‐
26] and Laboratory Diagnosis Programme of EMP, nested polymerase chain reaction (PCR) was performed on each case simultaneously to amplify the unique region of 18S ribosomal RNA (rRNA) gene of
Plasmodium falciparum,
Plasmodium vivax,
Plasmodium malariae and
Plasmodium ovale and to distinguish whether the
Plasmodium infection was single or mixed. Primer details of nested PCR and their reaction cycle were showed in Additional file
2.
The Plasmodium species was judged according to detection at the YPMDRL as follows: (1) the case was determined as negative when the microscopic examination and PCR testing were both negative for Plasmodium infection; (2) when Plasmodium infection was only detected by either microscopic examination or PCR testing, the final species was decided by the result of positive method; (3) PCR testing was used to determine the Plasmodium species when there was a difference between the microscopic examination and the PCR testing results. The re-testing examinations were performed at the YPMDRL, and determination results were provided once a month to the county-level laboratories to revise their initial diagnoses.
Primary index and statistical analysis
The database and information, which included the type of malaria, location of initial diagnosis and reporting, source of malaria infection, and re-testing results by the YPMDRL. SPSS 16.0 software (SPSS, Inc., Chicago) was used to analyse the number of malaria cases initially diagnosed, number of re-tested diagnoses, and so on. Among these analyses, the number of re-tested diagnoses (PTD) was counted, which may be re-tested by both microscopic examination and by PCR testing at same time [
21,
23‐
26] or only re-tested by one of the two detection methods. Test species accuracy (TSA) refers to the consistency between the
Plasmodium species identified by YPMDRL and the species initially diagnosed by county-level laboratories, for assessment the accuracy of the initial diagnosis. The Chi-square test was performed to test the correlation on the re-testing rate or the consistency rates of two laboratories (i.e. the accuracy rate of the initial diagnosis) in different years. The significance level was 0.05, and ArcGIS 10.1 (Environmental Systems Research Institute, ESRI) was used to map the malaria cases.
Discussion
The China Malaria Diagnosis Reference Laboratory Network was established in 2012 and aimed to regulate China's provincial-level laboratories to re-testing examination of the malaria cases initially diagnosed by county-level laboratories. The recommended re-testing methods for
Plasmodium infection included microscopic examination and/or genetic testing [
24]. The genetic testing method was used to differentiate
Plasmodium species when microscopy could not confirm various species. Snounou’s nested-PCR method [
24], with high specificity and sensitivity was usually adopted to detect
Plasmodium infection [
27]. On account of the fact that the nested PCR amplification products could not only be directly used for DNA sequencing and analysing of the gene polymorphism of
Plasmodium, Yunnan Province still applied the nested PCR based on Snounou's method to re-test
Plasmodium infection during EMP, even though the real-time fluorescence quantitative PCR was faster [
28,
29] and more sensitive [
30] in genetic diagnosis of malaria. In order to better control the internal quality evaluation, the YPMDRL persists in the strategy that every malaria case originally diagnosed must be simultaneously re-tested by microscopic examination and by gene testing [
21,
24]. By the end of 2014, the consistency of the two methods conducted by the YPMDRL reached 97.4% [
21]. The 95.6% (2742/2,869) re-testing rate for malaria cases initially diagnosed in Yunnan province in nearly six years was significantly higher than 45.5% from the WHO report 2018 [
8].
Moreover, the practice of identifying
Plasmodium species based on gene testing could also help microscopists in Yunnan Province to correct their understanding of the unfamiliar morphology of
Plasmodium. In recent years,
P. falciparum infection cases imported from Africa have increased in Yunnan Province [
31]. Once these falciparum malaria cases imported from Africa cause a clinical outbreak, the
Plasmodium morphology in peripheral blood often appears as mature trophozoites. This morphology of
P. falciparum was easily confused with the immature trophozoites of
P. vivax before 2016 in Yunnan Province [
21,
32], with the error rate of 55.6% in 2016 (Additional file
3). Moreover, a lack of familiarity with the morphology of
P. ovale and
P. malariae is also another major cause leading to incorrect identification of
Plasmodium various species in county-level laboratories in Yunnan Province (Additional file
3).
Since 2013, 20 cases of
P. ovale infected from Africa [
27] and
P. malariae infected from Myanmar were identified during the re-testing at the YPMDRL [
31], but all these cases were misdiagnosed as
P. vivax in county-level laboratories. It is possible that microscopists ignored to observe the morphological changes of erythrocytes parasitized by non-falciparum
Plasmodium species. It was observed [
33‐
36] that the different growth stages of
Plasmodium, such as trophozoites, schizonts and gametocytes with different proportions, appeared simultaneously in the peripheral blood of patients with
P. vivax,
P. malariae and
P. ovale. However, the morphology changes of erythrocytes were significantly different between three non-falciparum
Plasmodium species, which the erythrocytes parasitized by
P. vivax were generally enlarged, those parasitized by
P. malariae were generally smaller than normal ones, and those parasitized by
P. ovale were oval in shape [
32] and contained obvious Schüffner's dots [
37].
Nevertheless, the species accuracy rate of malaria initially diagnosed in Yunnan Province remained above 92% (Additional file
2) from 2013 to 2018, which is significantly higher than the 62–66% accuracy rate that Yin et al. reported in China in 2015 [
20]. A potential reason for this could be the different microscopists selected for assessment between the two studies. In the Yin et al
. study, those personnel were not always specialized malaria microscopists, but testers who could undertake the pathogenic testing for multiple parasitic diseases at the same time except for
Plasmodium. Retention of multiple testing skills may have limited microscopists in mastering
Plasmodium identification. Moreover, there were significant regional differences in the species accuracy rate between prefectures in Yunnan Province. Generally, in those prefectures that are traditional malaria endemic areas with a high number of diagnosed malaria cases, such as Dehong and Baoshan, the species accuracy rate was higher than the prefectures that are non-traditional malaria endemic areas with a low number of diagnosed malaria cases, such as Kunming and Zhaotong. The phenomenon that laboratories in areas with more diagnosed malaria cases have a stronger ability to precisely identify
Plasmodium species is consistent with the findings of studies conducted by Abanyie et al. [
18], Loomans et al. [
11], Alemu et al
. [
12], and Mukadi et al
. [
38]. It is suggested that continuous practice is advantageous in maintaining the microscopic skills required to examine
Plasmodium at a high level and stable state. In the next stage, it is necessary to strengthen the training of microscopic examination skills for
P. malariae and
P. ovale, and also how to distinguish between
P. ovale curtisi and
P.ovale wallikeri on the basis of gene polymorphism information analysis.
Plasmodium density was considered an effective indicator in predicting the severity of a malaria attack [
10,
39], but the China Malaria Diagnostic Criteria does not require
Plasmodium density to be measured during malaria diagnosis [
22]. Therefore, the
Plasmodium density was not collected during the initial diagnosis of malaria cases in Yunnan Province from 2013 to 2018, and the
Plasmodium density data was not also analysed in this paper. Additionally, although the organizer of EMP required YPMDRL to collect the every blood samples of malaria cases initial diagnosed and re-test their
Plasmodium infection at the same time, there were still some accidents, such as patients’ refusing to supply blood or too litter blood samples not enough to prepare the blood smears and dried blood spot specimens for re-testing, etc., that resulted in the missing (3.8%, 108/2869) or delayed collection (3.8%, 103/2742) for re-testing specimens. However, considering that the elimination malaria and consolidation its effect were a long-lasting monitoring plan, the few shortcomings in the programme design should be improved in the future. In addition, the introduced working model of implementation reference laboratory re-testing and application its’ results to revise the identification of malaria species in Yunnan Province are worthy of promotion in the other malaria elimination implementation areas.
Acknowledgements
We appreciate the support from the Centers for Disease Control and Prevention in states/cities and counties such as Dehong, Baoshan, Kunming, Pu'er, Lincang, Dali, Nujiang, Lijiang, Xishuangbanna, Yuxi, Chuxiong, Honghe, Zhaotong, Diqing, Qujing, and Wenshan.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.