This review found variation in methodologies and criteria used for scrub typhus indirect immunofluorescence assays (IFA), raising questions about which IFA threshold should be used to diagnose acute scrub typhus infection. These conclusions should be treated with caution as the analysis and included studies were unable to provide the strong evidence-base required to draw such conclusions.

To determine the evidence for scrub typhus indirect immunofluorescence assay (IFA) methodologies and criteria used for a positive test result.

PubMed was searched from 1963 (when the IFA test for diagnosis of scrub typhus was first described) to June 2005. The search terms, which were reported, included a diagnostic filter. Reference lists were screened to identify additional studies. Only studies published in peer-reviewed journals were included. Language restrictions do not appear to have been applied.

Studies in humans that used IFA to detect antibodies to Orientia tsutsugamushi (O. tsutsugamushi) were eligible for inclusion. Inclusion criteria were not defined further in terms of the participants, outcomes or study design. The included studies were case reports (clinical descriptions of at least one typhus case for which IFA was used for diagnosis), cause of fever studies (IFA used for diagnosis of scrub typhus), seroepidemiologic studies (IFA used to determine scrub typhus positivity, but not used to diagnose illness) and assay development and diagnostic accuracy studies (IFA used as reference standard against which other techniques were compared). The studies were performed in 12 different countries, mostly in Japan, Thailand, Malaysia and China. The studies were mainly conducted in the population that was exposed to O. tsutsugamushi, although some were conducted in travellers, patients in which samples were sent elsewhere for testing, or were conducted in more than one country.

The authors did not state how the papers were selected for the review, or how many of the reviewers performed the selection.

Studies were assessed according to the following five criteria: referencing of IFA methodology, description of antigens, reporting of antibody isotype, reporting of at least one criterion for a positive result, and justification of criteria for positive test result. The studies assigned a score of 1 for each quality item fulfilled, giving a total possible score of 5. Studies that detailed IFA as a novel method, that determined criteria for seropositivity by IFA, or that used IFA experimentally, were not subject to the quality assessment.

The authors did not state how many reviewers performed the quality assessment.

Data were extracted on prevalence by dividing the number of typhus antibody-positive patients (from seroepidemiology studies) or patients with confirmed cases using IFA (from diagnostic studies) by the total number of patients tested. Prevalence was not calculated for case reports or case-control studies. The data were analysed on the basis of country of exposure rather than the country where IFA was performed. Data on the antibody cut-off titre used were also extracted.

One reviewer extracted the data using a data extraction from, which a second reviewer then checked. Any disagreements were resolved through consensus or by referral to a third reviewer.

Multiple linear regression and analysis of variance were used to investigate the association between prevalence of O. tsutsugamushi infection, antibody cut-off titre, year of study publication and study population. The Kruskal-Wallis test was used to test for independence of the cut-off titre and the purpose of IFA. The nptrend test was used to conduct trend analysis of quality scores.

A total of 109 publications reporting 123 studies (14 publications each reported two substudies) were included in the review: 65 case reports, 28 seropositivity studies and 30 diagnostic accuracy studies.

Eight studies were excluded from the quality assessment. Cause of fever, seroepidemiology studies and diagnostic accuracy studies all scored a median of 3 (interquartile range, IQR: 2 to 4); case-studies were of poorer quality with a median score of 2 (IQR: 1 to 3). There was no overall improvement in quality over time.

Most studies used a defined cut-off titre rather than an increase in the titre in paired samples. The choice of positivity cut-off varied by country (p=0.012) and purpose of the IFA test. Only 25 of the 69 studies that stated a criterion for seropositive results justified this choice by citing a previous study. The most frequently cited study suggested a cut-off titre of 1:400 and at least a 4-fold increase in the titre in paired serum samples..

Prevalence data could only be extracted from 32 studies. Prevalence data were not reported in full but, where reported, ranged from 0.4 to 90.5%. There was a correlation between prevalence rate and positivity cut-off titre (r²=0.22, p=0.007). This correlation was higher in the seroprevalence studies (r²=0.33, p=0.008). The association was not significant for the fever studies (p=0.15) or diagnostic accuracy studies (p=0.14).

There is substantial variation in scrub typhus IFA methodologies and in the criteria used to define a positive result. Justification for the methodology and cut-off used were rarely reported. This raises questions about which (if any) IFA positivity cut-off value should be used to diagnose acute scrub typhus infection.

The review question was very broad and not reported clearly. Inclusion criteria were only defined in terms of the intervention; there was a very broad criterion for population (humans) and no details for outcomes or study design. The literature search had limited scope and included a diagnostic filter, and the review was restricted to published studies. It is therefore likely that relevant studies have been missed and the review may be subject to publication bias. Appropriate steps were taken to minimise bias in the extraction of data, but it is unclear whether such steps were also taken in the study selection and quality assessment processes. Although the quality of the studies was assessed, this referred more to the quality of reporting rather than the risk of bias. The types of included studies and analysis were unable to provide the strong evidence base needed to make recommendations about the appropriateness of different IFA thresholds. The authors' conclusions and recommendations for practice should therefore be interpreted with caution.

Practice: The data do not support the use of any single antibody titre to make a diagnosis of scrub typhus. Sequential samples should be used to assess the dynamics of the antibody response. Diagnosis for patients from endemic areas can only be made conclusively on the basis of at least a 4-fold increase in the titre in paired serum samples; diagnosis for travellers returning from endemic areas may be based on on a single positive titre.

Research: Researchers conducting future studies should perform preliminary studies in different geographical locations to determine the level of normal scrub typhus immunity in otherwise healthy populations, so that evidence-based positivity cut-off titres can be derived for the diagnosis of acute illness. There is a need for further research aimed at developing rapid and inexpensive non-serology-based means of diagnosis such as nucleic acid amplification or antigen detection.

UK Wellcome Trust.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.