IgM ELISA for leptospirosis diagnosis : a systematic review and meta-analysis ELISA IgM para diagnóstico de leptospirose : revisão sistemática

1 Programa de PósGraduação em Saúde Coletiva, Universidade do Extremo Sul Catarinense (UNESC). Av. Universitária 1105, Bairro Universitário. 88806-000 Criciúma SC Brasil. mir@unesc.net 2 Programa de PósGraduação em Ciências da Saúde, UNESC. Criciúma SC Brasil. 3 Laboratório de Epidemiologia, UNESC. Criciúma SC Brasil. IgM ELISA for leptospirosis diagnosis: a systematic review and meta-analysis


Introduction
Leptospirosis is a neglected infectious disease caused by spirochetes from the genus Leptospira.It constitutes the most widespread zoonosis and is emerging as a major public health problem with outcomes ranging from subclinical infections to fatal pulmonary hemorrhage and Weil´s syndrome 1 .
Leptospirosis has a broad geographical distribution, occurring in both rural and urban areas of tropical, subtropical and temperate regions.The disease outbreaks in developed countries are usually associated with occupational exposure, tourism or sporting events 1 .
Leptospirosis is transmitted by contact of abraded skin or mucous membranes with water or soil contaminated with urine from reservoir animals, such as rodents 2 .More than 500.000cases of severe leptospirosis are reported each year, with mortality rates exceeding 10% 3 .A new global estimate estimates that the overall annual incidence is 1 million cases and 60,000 deaths 4 .
The microscopic agglutination test (MAT) is most often used as a reference test 5 .Standard tests are tedious, laborious and require well-equipped laboratories with experienced staff and are therefore restricted to a few centers.Because the initial presentation of leptospirosis may be difficult to discern from other infectious diseases, rapid and accurate diagnosis is essential to prevent the progression of the more severe form of the disease, particularly in developing countries 2 .
Traditional serological methods, such as the ELISA, are widely used to diagnose leptospirosis.Antileptospires IgM may be detected 4 to 5 days after the onset of symptoms, before detection of IgG and agglutinating antibodies, and persist at least 5 months in patients 6 .ELISA can be performed with minimal training and typically provides results in 2-4 hours.The aim of this study was to perform a systematic review and meta-analysis of the literature to verify the accuracy of the IgM ELISA for leptospirosis diagnosis.

Methods
All methods for analysis, inclusion/exclusion criteria, data extraction and quality assessment were specified in advance.It was performed a systematic review according to a prospective protocol using PRISMA-statement guidelines 7,8  This sensitive filter was created by combining three filters to identify diagnostic studies via the Boolean operators "OR" and "AND".The search was limited to human studies and had no language restrictions.Reference lists of all available primary studies were reviewed to identify additional relevant citations.The complete search strategy is available on request.
Abstracts/titles identified from the search were screened by two reviewers.Disagreements about study inclusion or exclusion were initially solved by consensus, and if agreement was not possible, they were arbitrarily resolved by a third reviewer.
Cross-sectional and cohort studies, prospective and retrospective, which evaluated IgM enzyme-linked immunosorbent assay (Elisa) in Leptospirosis diagnosis were included.Studies that used the index test IgM Elisa to diagnose leptospirosis in patients were analyzed.The diagnostic reference standard was the result of the MAT with confirmation based on the result on the same serum sample as used for the index test.Therefore, the primary outcome analyzed was the presence of Leptospirosis.
It was extracted data on the studies, patients and test characteristics using a standardized form.Data were abstracted as 2 x 2 tables regarding IgM Elisa vs MAT in leptospirosis diagnosis (positive vs negative by cut-off).It was also calculated the sensitivities, specificities, and Odds Ratio diagnostic (DOR).Studies that lacked the data needed to construct 2 x 2 contingency tables were excluded.The assessment of non-Englishlanguage articles was performed independently following translation (if necessary).Any disagreement was resolved by consensus for studies published in all languages.Final inclusion or exclusion was made with reference to a selection criteria checklist.
Disagreements about study inclusion or exclusion were initially solved by consensus, and if agreement was not possible, they were arbitrarily resolved by another reviewer.The agreement statistics among reviewers were computed.
The methodological quality assessment for diagnostic accuracy was performed according to criteria from the Quality Assessment of Diagnostic Accuracy Studies (QUADAS 2) 9 .QUADAS-2 is designed to assess the quality of primary diagnostic accuracy studies, and it consists of four domains: patient selection, index test, reference standard, and flow of patients through the study and timing of the index tests(s) and reference standard "flow and timing".Signaling questions are included to help judge the risk of bias 8 .The Quality assessment of studies was independently performed using the Review Manager 5.2 software 10 .
The rates were calculated as true positive (TPR, sensitivity), false positive (FPR, 1 -specificity), true negative (TN) and false negative (FN) 11 .If any cell containing "0" was present in the contingency table, 0,5 was added to each cell to facilitate the calculations; if the study contained two cells with "0", the study was excluded from the analysis 12 .
Bivariate analysis was used to calculate pooled estimates of sensitivity, specificity, and DOR in addition to 95% confidence intervals (CIs) for the summary estimates 13 .The bivariate model preserves the 2-dimensional nature of diagnostic data by analyzing the logit transformed sensitivity and specificity of each study in a single model and considers both within-study and between-study variability, in contrast to the Littenberg and Moses method that departs from a fixed effects model 14 .To detect cut-off threshold effects, the relationship between sensitivity and specificity was evaluated by the Spearman's correlation coefficient.Pooled estimates were only calculated for studies showing sufficient clinical and statistical homogeneity.I 2 or Q tests (commonly used in meta-analysis) are not recommended for assessing statistical homogeneity in diagnostic reviews because they do not consider the association between sensitivity and specificity 15 .The DOR can relate to different combinations of sensitivities and specificities and describes the odds of the positive test resulting in participants with the disease compared with the odds of a positive test resulting in those without disease.A single diagnostic odds ratio corresponds to a set of sensitivities and specificities depicted by the SROC.It can change according to the threshold and to the ROC curve used to define an abnormal examination resulted in the expected trade-off between sensitivity and specificity.
A summary receiver operating characteristic curve was generated using data from all thresholds using the Littenberg and Moses method.Additionally, the area under the curve (AUC) can summarize the inherent capacity of a test for discriminating a diseased from a non-diseased subject.Accurate tests usually have AUCs close to 1, and poor tests usually have AUCs close to 0.5 16 .Sensitivity analyses were performed to assess excluding studies with a high risk of verification bias according to QUADAS 2. To analyze publication bias, inverted funnel plots of the logarithmic odds ratio (OR) of individual studies were plotted against sample size 15 .
The statistical analysis was performed with the software Stata 11 17 , Meta-DiSc® 18 (version 1.4), and Review Manager 5.2 10 .

Results
A total of 545 studies were identified: 510 studies were identified using the database search and 35 additional records were identified through other sources.Seventy-nine full-text articles were retrieved; 27 were excluded after further scrutiny.Fifty-two primary studies, including 10,775 serum samples, met the criteria for inclusion and were included in the meta-analysis 19-69 (Figure 1).
Details of the participants and interventions are summarized in Table 1   . Most tudies were prospective, except for two 41,44 .
The robustness of the results was tested by repeating the analysis using a different statistical model (random effects model).Some studies were identified as outliers, and one re-analysis was performed without them.However, no significant difference was found in the sensitivity or specificity; therefore, those papers were not excluded from the meta-analysis.All 52 studies selected were included in the meta-analysis.Statistical analyses were performed on both the acute and unspecific phase and only the acute phase.Analysis with excluding particular studies with high risk of bias 48,65 in relation to the index test were conducted, and because there was no significant change they were maintained the meta-analysis.
SROC curves were constructed due to heterogeneity in the DOR.The AUC for the ROC curve was estimated by a trapezoidal rule 95.The resulting summary ROC curves are shown with operating points for sensitivity and specificity.The AUC was 0.960 in all studies and 0.952 in the acute phase respectively (Figure 4).
Covariable-type studies were separated into prospective and retrospective design, and the meta-regression analysis indicated no association between type of studies and outcome (P = 0.32).
Begg's funnel plot and Egger's test were performed to assess the publication bias of the literature in all comparison models.The shape of the funnel plot reveals any evidence of obvious asymmetry.Then, the Egger's test was used to provide statistical evidence of funnel plot symmetry for total phase (P for bias = 0.001) and acute phase (P for bias = 0.008), indicating publication bias (data not shown).

Discussion
In summary, this systematic review showed that IgM ELISA in all phases had a sensitivity of 0.86 and specificity of 0.84, whereas the acute phase had a sensitivity of 0.90 and specificity of 0.91.
The results showed that IgM ELISA could be useful as a screening and a confirmatory test, especially in regions with small laboratories that have difficulty performing other techniques such as MAT.
A recent systematic review included 35 studies up to 2010 and analyzed ELISA (IgM, IgG and IgA).In the present study, 55 studies with IgM only were included and analyzed the accuracy of IgM in the acute phase of the disease.We found a higher sensitivity compared to IgM results Signorini et al. 71 , 86 versus 80%, respectively.
It was found high heterogeneity between studies.It is expected in meta-analyses of diagnostic test accuracy because it comes from observational studies, study designs and different cutoff points.This high heterogeneity was also observed in the meta-analysis performed by Signorini et al. 71 .A rapid diagnostic test provides a quick test result but does not indicate an early test.The ideal rapid test should have high accuracy, be easy to perform, interpret, inexpensive, and stable and give the result within 2 hours 70 .
There are two phases of Leptospira infection: (1) between 3-7 days or acute septicemic phase with nonspecific symptoms such as myalgia and headache.The leptospires are detectable in the blood stream, decrease until 15 days 72 and (2) the start in the second week after the onset of symptoms, and the antibodies usually persist for several months 6 .During this phase, leptospires are eliminated from the blood stream as IgM antibodies increase 73 .
The rapid test depends on the detectable presence of anti-Leptospira antibodies already presented during the acute phase of the disease 74 .Molecular tests that detect the causative agent can be confirmed during the first 5 days after the onset of the disease 75 .It is very important that a test be rapid and sensitive, because the earlier the diagnosis the faster the treatment decision.
Whereas molecular tests, such as the polymerase chain reaction (PCR), that demonstrate the presence of the causative agent in a clinical sample mainly during the first 5 days after the onset of the disease (DPO), serological tests depend on the accumulation of detectable amounts of anti-Leptospira antibodies in the late acute to convalescent samples [74][75][76] .
Rapid diagnostic tests should ideally be accurate, simple to use, relatively inexpensive, easy to interpret, stable under extreme conditions, require little or no processing, and give the results within 1-2 hours 70 .Again, it is very important that a test be rapid and sensitive, because the earlier the diagnosis the faster the treatment decision.
Often, an early diagnosis or reference standard is employed in referral centers where confirmation is performed by experts.The rapid diagnosis is highly useful at the peripheral facilities and might be integral for early outbreak warning and useful for monitoring outbreaks if a rapid unusual accumulation of cases might provide an early alert, provided that specimens are collected, transported, and stored in an adequate manner 76 .
This review, which included retrospective and prospective studies, had the following limitations: i) high heterogeneity found between studies; ii) use of selected samples and the choice of case definition may be a source of bias; and iii) it is a misunderstanding that rapid tests are easy and therefore do not require experience; iv) it may reflect population-related differences, such as past exposure to leptospirosis, exposure to environmental leptospires, or infection with other infectious agents.
In conclusion, in the meta-analysis, the diagnosis of leptospirosis was ascertained by definite clinical criteria and standard MAT criteria.Also, IgM ELISA is sufficiently sensitive for use as an initial screen for leptospiral infections.The IgM ELISA showed higher sensitivity (84%) and specificity (91%) in the diagnosis of acute leptospiral infection and can be used as a rapid test for the detection of the disease, therefore improving the prognosis of patients and decreasing the lethality of leptospirosis.

Figure
Figure 2. Results of the evaluation of each study according to QUADAS 2.

Figure 4 .
Figure 4. Summary receiver operating characteristic curves.A: all studies and B: acute phase.

Figure 1 .
Flow diagram of the study selection process.

Table 1 .
Characteristics of the primary diagnostic studies.

2 .
Results of the evaluation of each study according to QUADAS 2. Forest plot of sensitivity (A) and specificity (B) of the all studies included in this review.