Diagnostic value of loop‐mediated isothermal amplification assay for hand, foot, and mouth disease

Abstract Background Nowadays, hand, foot, and mouth disease (HFMD) has a significant negative impact on children's health, especially in the Asia‐Pacific region. Loop‐mediated isothermal amplification assay (LAMP) is a highly efficient and convenient novel tool. However, its diagnostic accuracy for HFMD is still not clear. Therefore, we conducted a meta‐analysis in order to evaluate the potential of LAMP assay for the diagnosis of HFMD, in which the reference standard was polymerase chain reaction (PCR). Methods A protocol was predetermined (CRD42020212882) in PROSPERO. We retrieved seven databases including PubMed for relevant studies published before October 2020. Articles were included if they compared the diagnostic efficiency of LAMP with PCR for HFMD through detecting clinical samples which was more than 15. Statistical analysis was performed by STATA 15.1 software. Risk of bias and applicability were assessed using Quality Assessment of Diagnostic Accuracy Studies. No funding was used for the study. Results A total of 18 retrospective studies including 2495 samples from China were finally included. Reference standards of them included RT‐PCR and non‐RT‐PCR. The merged sensitivity and specificity with 95% confidence interval (95% CI) were 1.00 (0.97–1.00) and 0.97 (0.88–0.99), respectively. The pooled PLR, NLR, and DOR with 95% CI were 11.17 (5.91–21.11), 0.05 (0.03–0.09), and 538.12 (183.17–1580.83), respectively. The AUC of SROC was 1.00 (95% CI: 0.99–1.00). Conclusion In conclusion, our research revealed high sensitivity and specificity of LAMP in diagnosing HFMD. However, more high‐quality research is required to prove this conclusion.


| INTRODUC TI ON
Hand, foot, and mouth disease (HFMD) is a common infectious disease in children caused by Enterovirus A, in which the most generally reported genotypes are Enterovirus A71 (EV-A71) and Coxsackievirus A6, A10, and A16 (CVA6, CV-A10, and CV-A16). 1 It is common in preschoolers. In China, children below the age of 3 are most susceptible to HFMD. 2 HFMD is characterized by macular papules and herpes on the palms, soles, oral mucosa, and buttocks, with or without fever. 3 It is prone to central nervous system damage and severe cases of brainstem encephalitis, leading to neurogenic pulmonary edema, pulmonary hemorrhage, and even death. 4 Timely and correct diagnosis is of great significance for its treatment.
The traditional laboratory diagnostic methods of HFMD are mainly isolation culture and serological detection. However, because of the need for a particular culture medium, time-consuming and complex experimental conditions, and comparatively low detection rate leading to late diagnosis, isolation culture has not been routinely used in HFMD diagnosis. The practical value of serology in the diagnosis of early HFMD infection is limited since serological diagnosis is time-consuming. 5 It is usually used as a retrospective diagnosis and epidemiological investigation to provide the basis for the formulation of prevention and control measures. 6 In recent years, with the rapid development of molecular biotechnology, real-time fluorescent quantitative polymerase chain reaction (PCR) has become a routine method for laboratory diagnosis of HFMD. 7,8 However, a lot of money is required in the early stages to meet the needs of the environment, personnel, and instruments, so it is difficult to detect this in a primary laboratory. With the shortcoming of requiring costly precise equipment, PCR may not be a good method to diagnose HFMD in the basic clinical place, especially for developing countries. 9 Therefore, it is necessary to find a rapid, economical, simple, sensitive, and specific tool to diagnose HFMD.
Loop-mediated isothermal amplification (LAMP) is a newly emerged nucleic acid amplification method, which relies upon DNA polymerase with strand replacement ability and two pairs of primers, which can identify six regions of the target sequence. 10 The whole detection reaction can be completed only at 65℃ for 1-2 h, and the results can be judged by the naked eye. Compared with other detection methods, this method is simple to operate and the equipment used is not expensive. 11 The most prominent feature of LAMP is that it allows instant diagnosis. 12 It has a widespread application in the diagnosis of infectious diseases such as malaria and trypanosomiasis. 13 Moreover, it has been regarded as a potential alternative to PCR in the laboratory. 14 However, the diagnostic accuracy of LAMP for HFMD diagnosis remains unclear. Hence, we performed a pooled-analysis to evaluate the overall efficacy of LAMP in diagnosing HFMD. PCR was served as the reference standard.

| Search strategy
We identified all the studies published before October 2020 by systematically searching Embase, PubMed, Cochrane Library, Web of Science, Wan Fang Data, SinoMed, and the Chinese National Knowledge Infrastructure (CNKI) databases using the following strategy as follows: ("Hand, Foot, Mouth Disease" OR "Hand, Foot and Mouth Disease" OR "hand foot and mouth disease" OR "hand foot mouth disease") AND ("LAMP" OR "isothermal amplification loop-mediated" OR "loop-mediated isothermal amplification" OR "loop mediated isothermal amplification").
Then, we inspected the bibliographies of all the publications to complement the retrieval.

| Study selection
Two review authors carefully reviewed all the retrieved articles and selected the articles according to the inclusion and exclusion criteria established in advance. If they had different opinions on some articles, they discussed them with a third review author until they were in agreement.

| Inclusion and exclusion criteria
Articles were included in the analysis if (1) clinical samples of human were analyzed; (2) they compared the diagnostic accuracy of LAMP with PCR for HFMD; (3) PCR was served as the reference standard; (4) the sample size was more than 15; (5) the generated data sufficed to construct two-by-two tables including true positive (TP), false positive (FP), false negative (FN), and true negative (TN) for working out the sensitivity, specificity, and likelihood ratios.
Criteria for excluding the studies were as follows: (1) Duplicate publications; (2) Abstracts, case reports, letters, reviews, and editorials.

| Data extraction
The following information was extracted individually by two review authors from the articles selected for inclusion: Publication information, sample size, specimen type, gold standard, and virus detected type. Data for 2·2 tables (TP, FP, FN, and TN) were extracted. When inconsistencies were encountered, the two review authors resolved them by reaching an agreement with a third one.

| Quality assessment
Relying on the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) guidelines, 16

| Subgroup meta-analysis
Subgroup analysis was carried out depending on whether the standard reference was reverse transcription polymerase chain reaction (RT-PCR) and on the types of samples, respectively.

| Literature searching result
A total of 422 studies were identified from which 281 duplicates were discarded. After screening their titles and abstracts, 119 articles were excluded, and 22 articles were left to browse the full-text. Finally, three articles were removed because of the lack of gold standard or due to insufficient data to construct a 2×2 table.
Another article was discarded since the index test was combined with additional techniques. The remaining 18 studies 9,17-33 which were eligible were included to conduct pooled-analysis. The flow diagram is shown in Figure S1.

| Quality assessment
Most studies had a low risk of bias in most domains. And they all had low concern for applicability. However, 11 (61.1%) of the studies were considered to have a high risk of bias in the patient selection domain due to the inclusion of participants with confirmed diagnoses. Risk of bias of another two studies 17,20 was assessed as high in the index test domain because they did not pre-specify a threshold. Reference standard results of Xia's study 17 were not interpreted without knowledge of the results of the index test. Therefore, this study was considered as high risk in the reference standard domain. The results of independent studies and the overall results are shown in Figures 1 and 2, respectively.

| Threshold effect
"Shoulder-arm" distribution of scatter plots did not exist in the SCOR curve (Figure 3). The proportion of heterogeneity likely due to threshold effect was 0.17. There was probably no threshold effect in the included studies.

| Merge analysis results
To assess the diagnostic efficacy of LAMP assay using a random-

| SROC curve
The calculated AUC was 1.00 (95% CI: 0.99-1.00) as presented in Figure 3, indicating the high diagnostic accuracy of LAMP for HFMD.

F I G U R E 3 Summary receiver operating characteristic (SROC) curves of HFMD diagnosed by LAMP
F I G U R E 4 Forest plots for the pooled sensitivity and specificity of LAMP

| Meta-regression analysis and subgroup analysis
It could be observed in Figure 5 that only the reference standard variable had statistical significance for sensitivity (p < 0.001). The results of the subgroup analysis were shown in Table 2.

| Publication bias
As shown in Figure 6, significant publication bias was discovered in Deeks' funnel plot (p = 0.00).

| DISCUSS ION
Hand, foot, and mouth disease poses a heavy disease burden mainly in Asia and the Pacific Rim, especially among children under the age of 5. In some severe cases, it can result in death due to fatal cardiopulmonary and neurological complications. 34 Thus, it is of great significance to master rapid and effective diagnostic methods for the treatment of HFMD. Our study concentrated on investigating the diagnostic value of the LAMP assay for HFMD.
After merging the results of 18 articles, we gained the merged sensitivity, specificity, PLR, NLR, and DOR, which were 1.00, 0.97, 11.17, 0.05, and 538.12, respectively. DOR was so high that it was obvious that LAMP had a high diagnostic value for HFMD.
Nevertheless, the following items also proved the point. First, the SROC curve was far away from the middle diagonal and lay close to the upper left corner. At the same time, its AUC was 1.0. Second, NLR < 0.1 and PLR > 10, suggesting that LAMP was reliable in judging whether a person had HFMD. Third, high sensitivity and specificity suggested that the possibility of misdiagnosis and missed diagnosis were both low. Nevertheless, it was found that higher sensitivity and specificity existed in the group that used both pharyngeal swabs and fecal samples than the one that used just one of them, which indicated that various samples might improve the diagnosis value. Additionally, compared with the group that took RT-PCR as F I G U R E 5 Meta-regression results a reference standard, the non-RT-PCR group had higher sensitivity and specificity, which indicated that using non-RT-PCR as a reference standard might overstate the diagnosis accuracy. results. That is to say, we were supposed to draw a conclusion based on the merged results with prudence.
Our study still had some limitations. First, the articles we included all originated in China, so the lack of data from other countries might impact the results. Second, some studies were considered to have a high risk of bias in certain domains, which might increase the risk of bias. Studies that did not refrain from a case-control design might change the diagnostic accuracy. 38 In Xia's and Zhao's study, they did not pre-specify a threshold for the index test, which might result in measurement bias. In Xia's study, interpretation of reference standard results with knowledge of the index test results might also increase the measurement bias. Third, the sources of high heterogeneity were not figured out completely. Besides, it was unclear whether unpublished negative results affected the existing results significantly.
In conclusion, according to the present combined results, LAMP has high sensitivity and specificity for the diagnosis of HFMD. That is to say, it has a high diagnostic value for HFMD. With the characteristic of high efficiency and convenience, LAMP may become the main laboratory diagnostic tool for HFMD in the future. However, more high-quality research is required to prove this conclusion.

CO N FLI C T O F I NTE R E S T
The authors declare that there are no competing interests associated with the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Not applicable.