Evaluation of Brucella IgM and IgG ELISA Responses in Relation to Blood Culture Positivity in Pediatric Brucellosis

Abstract

Background

Class-specific Brucella IgM and IgG enzyme-linked immunosorbent assay (ELISA) responses according to blood culture status are incompletely characterized in children with confirmed brucellosis. We compared IgM and IgG ELISA findings according to contemporaneous blood culture status among children already classified as having confirmed acute brucellosis; diagnostic accuracy was not assessed.

Methods

This retrospective, single-center, cross-sectional observational study included 94 children aged 18 years or younger. Rose Bengal testing, standard tube agglutination testing (STAT), Brucella IgM and IgG ELISA, and blood culture were obtained at the same presentation. Serologic findings were compared by culture status, and factors associated with culture positivity were explored using multivariable logistic regression.

Results

Blood cultures yielded Brucella melitensis in 49/94 children (52.1%). IgM ELISA was positive in 93/94 (98.9%). IgG ELISA was positive in 70/94 (74.5%), equivocal in 4/94 (4.3%), and negative in 20/94 (21.3%). IgM values did not differ significantly between culture-positive and culture-negative groups (41.2 ± 15.8 vs 36.2 ± 12.6 NTU; p=0.096). IgG values were lower in culture-positive children (median 18.5 vs 21.0 NTU; Mann–Whitney U=789; p=0.018; rank-biserial correlation=0.284). After excluding equivocal results, IgG negativity was more frequent in culture-positive than culture-negative children (14/47 [29.8%] vs 6/43 [14.0%]), but not significantly (Fisher exact p=0.082). Fever with arthralgia, elevated C-reactive protein, and elevated aspartate aminotransferase remained associated with culture positivity in the exploratory model.

Conclusion

Among children with confirmed acute brucellosis, IgM ELISA was positive in nearly all patients, while continuous IgG values were lower in culture-positive children. These findings describe associations with contemporaneous culture status and should not be interpreted as antibody kinetics or diagnostic-performance estimates. ELISA values should not be used alone to predict blood culture positivity or determine whether blood culture should be obtained.

Plain Language Summary

Brucellosis is an infection that can pass from animals to people, often through contact with infected animals or consumption of unpasteurized dairy products. In children, symptoms can be nonspecific, so several laboratory methods may be used together.

We studied 94 children who already had confirmed acute brucellosis and compared their antibody test results according to whether their blood culture was positive or negative. Blood culture detects the bacteria directly, whereas IgM and IgG ELISA tests measure antibodies produced by the immune system.

IgM was positive in almost all children. IgG results were positive in 70 children, equivocal in 4, and negative in 20. Children with positive blood cultures had lower IgG values on average, but antibody levels alone could not determine whether a blood culture would be positive.

In children with compatible symptoms and epidemiological exposure, a negative IgG result may still occur in confirmed acute brucellosis and should therefore be interpreted together with other clinical and laboratory findings. Blood culture and antibody testing provide complementary information, and neither IgM nor IgG level should be used alone to decide whether blood culture should be obtained.

Introduction

Human brucellosis remains an important zoonotic infection, with recent estimates suggesting more than two million cases annually.¹ Transmission commonly occurs through consumption of unpasteurized dairy products or direct contact with infected animals, and the disease remains endemic in Türkiye, particularly in livestock-raising regions.^2,3 In children, brucellosis frequently presents with nonspecific constitutional and musculoskeletal manifestations, making laboratory confirmation central to timely recognition.²

Diagnosis requires integration of epidemiologic exposure, compatible clinical findings, microbiological evidence, and serologic testing. Blood culture provides etiologic confirmation but has incomplete sensitivity because recovery is influenced by disease stage, intermittency of bacteremia, specimen volume, laboratory procedures, and previous antimicrobial exposure.^3–5 Rose Bengal testing and STAT remain widely used, while class-specific ELISAs provide additional information about IgM and IgG responses.^3–6

The humoral response to smooth Brucella species evolves during acute infection. IgM generally appears first, followed by IgG as the response matures; the relative contribution of the two antibody classes therefore depends partly on the timing of sampling.⁴ This temporal biology is clinically relevant because an IgM-positive/IgG-negative profile may occur during early infection, while persistent or renewed IgG responses may carry different meaning later in the disease course.^4,7,8

Previous pediatric studies comparing culture-positive and culture-negative brucellosis have focused mainly on clinical manifestations, inflammatory markers, transaminase abnormalities, and agglutination results.^9–11 IgM and IgG ELISA responses have been compared by bacteremia status in mixed-age cohorts,¹² but class-specific continuous ELISA values together with qualitative IgM/IgG patterns according to contemporaneous blood culture status remain insufficiently characterized in children with confirmed acute disease.

Accordingly, the primary objective of this study was to compare Brucella IgM and IgG ELISA responses according to contemporaneous blood culture status among children already classified as having confirmed acute brucellosis. Secondary objectives were to describe class-specific serologic patterns and to explore clinical and laboratory factors associated with culture positivity. Because the study included only confirmed cases and no brucellosis-negative comparator group, it was not designed to estimate ELISA sensitivity, specificity, predictive values, likelihood ratios, or overall diagnostic accuracy.

Materials and Methods

Study Design and Setting

A retrospective, single-center, cross-sectional observational study was conducted at the Pediatric Infectious Diseases Outpatient Clinic of Health Sciences University Gazi Yaşargil Training and Research Hospital, Diyarbakır, Türkiye, covering pediatric encounters from January 1, 2021 through August 1, 2022. The hospital serves as a referral center for southeastern Türkiye, where human brucellosis remains endemic. The study was reported in accordance with the STROBE statement for observational research.¹³

Participants and Study Framework

Electronic medical records and the microbiology laboratory information system were screened for patients aged 18 years or younger with newly diagnosed acute brucellosis during the study period. Reported symptom duration did not exceed three weeks; however, exact duration within this interval was not consistently available in the retrospective records. Eligibility required Rose Bengal testing, STAT, Brucella-specific IgM and IgG ELISA, and blood culture obtained at the same index presentation. All eligible patients were included; no a priori sample-size calculation or prospective enrollment was performed, yielding an analytic sample of 94 children.

A confirmed case required (1) compatible clinical findings documented at presentation, including constitutional and/or musculoskeletal manifestations compatible with acute brucellosis; (2) recognized epidemiological exposure; and (3) either recovery of B. melitensis from blood culture or a STAT titer ≥1:160. Rose Bengal testing was treated as a screening method and was not accepted as the sole case-defining criterion. All 45 culture-negative participants had STAT titers ≥1:160 and therefore fulfilled the laboratory component of the case definition independently of blood culture. Blood culture positivity contributed to case confirmation in some participants and was subsequently used as the grouping variable for this within-case comparison; culture-negative children were not considered disease-negative controls.

Specimen Collection and Laboratory Procedures

Venous blood was drawn aseptically at the index presentation. For blood culture, samples were inoculated into BACT/ALERT bottles (bioMérieux, Marcy l’Étoile, France) and incubated in the BACT/ALERT system for up to 10 days; per the laboratory’s standing protocol, incubation was not extended beyond this 10-day window. Positive bottles were processed, and isolates were reported as B. melitensis following identification with the VITEK 2 automated system (bioMérieux, Marcy l’Étoile, France). Biotyping and molecular subspecies confirmation were not part of the routine workflow and were not undertaken on the study isolates.

Serologic assays were completed on the day of sampling. The Rose Bengal screen used 30 μL of patient serum combined with 30 μL of a commercial antigen suspension (bioMérieux, Marcy l’Étoile, France) on a glass slide; the mixture was continuously agitated, and macroscopic clumping visible at the 4-minute read was scored as positive. STAT was carried out with Brucella-WRIGHT antigen: serial twofold dilutions of serum were prepared and incubated overnight at 37 °C, with a titer ≥1:160 read as positive; the laboratory’s routine did not include endpoint dilution beyond 1:160, so titers above this level were recorded as “>1:160” without further quantitation.

Class-specific antibodies were quantified with the BRUCELLA ELISA IgM (reference M1006) and BRUCELLA ELISA IgG (reference G1003) kits from Vircell S.L. (Granada, Spain), with results expressed in NTU. The manufacturer-defined interpretive bands were applied: <9 NTU negative, 9–11 NTU equivocal, and >11 NTU positive. Both numerical IgM and IgG results and their qualitative patterns were included in the analyses. Equivocal results were retained as a separate category for descriptive reporting and were excluded from binary positive-versus-negative categorical comparisons. Serologic assays were performed on the day of index sampling, before final blood culture results became available. Because of the retrospective design, patient-level documentation was insufficient to reliably determine duplicate-testing status, formal blinding procedures, or additional internal-control details beyond the manufacturer-defined assay procedure.

Routine hematology was generated on a Sysmex XN-1000 analyzer using EDTA-anticoagulated whole blood. Clinical chemistry — including AST, ALT, and CRP — was measured on a Beckman Coulter AU5800 analyzer, with transaminases quantified by an enzymatic kinetic method tracking NADH absorbance and CRP quantified by immunoturbidimetry.

Clinical and Laboratory Variables

Recorded clinical covariates abstracted from the medical record were age, sex, household brucellosis contact (defined as a documented diagnosis in a parent or sibling living in the same household), livestock exposure, and consumption of unpasteurized dairy products. These epidemiological exposures were observed characteristics of the study population and were not individual mandatory eligibility criteria. The symptoms recorded at the index presentation—fever, arthralgia, malaise, sweating, weight loss, and low back pain—were captured in binary form; the composite “fever with arthralgia” required both symptoms to be documented. Hepatosplenomegaly was accepted as present when recorded on physical examination and/or on abdominal ultrasonography performed for clinical indications.

Laboratory covariates included complete blood count, AST, ALT, and CRP. Anemia was classified using age- and sex-specific hemoglobin thresholds reported by Braat et al.¹⁴ Thrombocytopenia was defined as platelet count <150 × 10⁹/L, elevated CRP as >6 mg/L, and elevated AST as >50 U/L. Calendar month of diagnosis was also recorded. Monthly case counts were summarized descriptively; because the study period did not provide equal observation time for every calendar month, these counts were not used for formal seasonal inference.

Statistical Analysis

Descriptive and univariable analyses were performed using IBM SPSS Statistics version 24.0 (IBM Corp., Armonk, NY, USA) and jamovi for the additional IgG analyses. Normality was assessed with the Shapiro–Wilk test. Continuous variables are reported as mean ± SD for approximately normal distributions and as medians where non-normal; categorical variables are reported as counts and percentages. Between-group comparisons by culture status used Student t test for approximately normally distributed continuous variables, the Mann–Whitney U-test for non-normally distributed continuous variables, and Pearson chi-square or Fisher exact test for categorical variables, as appropriate.

Positivity proportions are summarized with Wilson 95% confidence intervals. IgM and IgG qualitative patterns were described using cross-tabulations, with manufacturer-defined equivocal ELISA results retained as a separate descriptive category. Equivocal results were excluded from binary positive-versus-negative categorical comparisons. Continuous IgG values were compared by culture status with the Mann–Whitney U-test, and IgG negativity by culture status was evaluated with Fisher exact testing among samples with interpretable positive or negative IgG results. Unadjusted associations between categorical clinical and laboratory variables and culture positivity are reported as odds ratios with 95% confidence intervals.

An exploratory multivariable logistic regression was retained to examine factors associated with blood culture positivity. The six pre-specified candidate variables entered simultaneously were fever with arthralgia, elevated CRP, elevated AST, anemia, thrombocytopenia, and the continuous IgM ELISA value. The detailed IgG analyses were secondary serologic analyses and were not added post hoc to the pre-existing model solely on the basis of their univariable findings. All 94 participants had complete data for the six model variables. With 49 culture-positive events and six predictors, the model had approximately 8.2 events per predictor; therefore, model instability and overfitting remain possible. Multicollinearity was assessed using variance inflation factors (VIFs), which ranged from 1.05 to 1.36, indicating no problematic multicollinearity. The model was not internally or externally validated and should be interpreted strictly as an exploratory association model rather than a clinical prediction instrument. All tests were two-sided with α=0.05.

Ethical Approval

The study was reviewed and approved by the Institutional Review Board of Health Sciences University Gazi Yaşargil Training and Research Hospital (approval no. 165; September 13, 2024), and was conducted in accordance with the Declaration of Helsinki. The requirement for informed consent was waived because of the retrospective design and the use of fully anonymized clinical data.

Results

Participant Characteristics

Ninety-four children constituted the analytic sample: 52 boys (55.3%) and 42 girls (44.7%), with a mean age of 120.0 ± 55.3 months. Brucellosis in a parent or sibling living in the same household was documented in 55 patients (58.5%), sheep breeding in 67 (71.3%), and consumption of unpasteurized milk or dairy products in all 94. These were observed epidemiological characteristics of the local study population rather than individual mandatory eligibility criteria. Malaise was recorded in 84 (89.4%), arthralgia in 64 (68.1%), and fever in 42 (44.7%). The full distribution of demographic, exposure, clinical, and laboratory characteristics is shown in Table 1. The raw monthly distribution of diagnoses is presented descriptively in Figure 1; because calendar months had unequal observation time within the study interval, these counts cannot support formal epidemiological inference regarding seasonality.

Table 1.

Demographic, Exposure, Clinical, and Laboratory Characteristics of the Cohort

Characteristic	n (%) or Mean ± SD
Sex
Male	52 (55.3)
Female	42 (44.7)
Age, months	120.0 ± 55.3
Brucellosis in a parent or sibling living in the same household	55 (58.5)
Sheep breeding	67 (71.3)
Consumption of unpasteurized dairy products	94 (100.0)
Symptoms
Malaise	84 (89.4)
Arthralgia	64 (68.1)
Fever	42 (44.7)
Sweating	28 (29.8)
Weight loss	18 (19.1)
Fever with arthralgia	17 (18.1)
Low back pain	15 (16.0)
Physical examination
Hepatosplenomegaly	13 (13.8)
Laboratory findings
Elevated AST	26 (27.7)
Elevated CRP	25 (26.6)
Anemia	21 (22.3)
Thrombocytopenia	10 (10.6)

Notes: Anemia was defined by age- and sex-specific hemoglobin thresholds;¹⁴ thrombocytopenia by a platelet count <150 × 10⁹/L; elevated CRP by >6 mg/L; and elevated AST by >50 U/L.

Abbreviations: AST, aspartate aminotransferase; CRP, C-reactive protein; SD, standard deviation.

Figure 1.

Descriptive monthly distribution of pediatric brucellosis diagnoses during the study interval. Bars show pooled calendar-month case counts. Because the study period did not provide equal observation time for all calendar months, the figure is descriptive and should not be interpreted as a formal analysis of seasonality.

Serologic Profiles and Blood Culture Results

Blood cultures recovered B. melitensis in 49 of 94 children (52.1%; Wilson 95% CI, 42.1–61.9%). Brucella IgM ELISA was positive in 93/94 (98.9%; 95% CI, 94.2–99.8%), negative in 1/94 (1.1%), and equivocal in none. Brucella IgG ELISA was positive in 70/94 (74.5%; 95% CI, 64.8–82.2%), equivocal in 4/94 (4.3%), and negative in 20/94 (21.3%). Among the 90 children with interpretable positive or negative IgG results, 19 had an IgM-positive/IgG-negative pattern, 70 had an IgM-positive/IgG-positive pattern, and one was negative for both IgM and IgG. No child had an IgM-negative/IgG-positive pattern. The four children with equivocal IgG results were IgM positive. Rose Bengal and STAT were each positive in 91/94 children (96.8%; 95% CI, 91.0–98.9%). Table 2 summarizes observed positivity proportions, and Supplementary Table S1 shows the complete class-specific serologic distribution including equivocal IgG results.

Table 2.

Observed Positivity Proportions of Laboratory Tests Within the Confirmed Acute Brucellosis Cohort

Method	Positive/total	Positive, %	95% CI
Brucella IgM ELISA	93/94	98.9	94.2–99.8
Brucella IgG ELISA	70/94	74.5	64.8–82.2
Rose Bengal test	91/94	96.8	91.0–98.9
STAT	91/94	96.8	91.0–98.9
Blood culture	49/94	52.1	42.1–61.9

Notes: Four IgG results were equivocal and are not included in the IgG-positive numerator; no IgM result was equivocal. Confidence intervals were calculated using the Wilson method.

Abbreviations: CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; STAT, standard tube agglutination test.

Continuous IgM values were numerically higher in culture-positive than culture-negative children, but the difference was not statistically significant (41.2 ± 15.8 vs 36.2 ± 12.6 NTU; p=0.096). Continuous IgG values were lower in culture-positive children (median 18.5 vs 21.0 NTU; mean ± SD, 15.3 ± 7.67 vs 19.1 ± 8.06 NTU; Mann–Whitney U=789; p=0.018; rank-biserial correlation=0.284). Four IgG results were equivocal (two in each culture group) and were excluded from the binary IgG-negative comparison. Among the remaining interpretable results, IgG negativity was numerically more frequent in culture-positive than culture-negative children (14/47 [29.8%] vs 6/43 [14.0%]), but this categorical difference did not reach statistical significance (Fisher exact p=0.082; Table 3). Thus, the statistically significant finding relates to the distribution of continuous IgG values and does not establish a clinically useful categorical threshold for bacteremia.

Table 3.

Brucella IgM and IgG ELISA Findings According to Blood Culture Status

Serologic Finding	Culture Positive (n=49)	Culture Negative (n=45)	p value
IgM ELISA, mean ± SD (NTU)	41.2 ± 15.8	36.2 ± 12.6	0.096
IgG ELISA, median (mean ± SD), NTU	18.5 (15.3 ± 7.67)	21.0 (19.1 ± 8.06)	0.018
IgG negative among interpretable results, n/N (%)	14/47 (29.8)	6/43 (14.0)	0.082

Notes: Normality was assessed using the Shapiro–Wilk test. IgM values were approximately normally distributed and compared using Student t test; continuous IgG values were non-normally distributed and compared using the Mann–Whitney U-test. Four equivocal IgG results (two per culture group) were excluded from the binary IgG-negative comparison, which used Fisher exact testing.

Abbreviations: ELISA, enzyme-linked immunosorbent assay; NTU, numerical test units; SD, standard deviation.

Factors Associated with Blood Culture Positivity

Age and sex distributions were similar between culture-positive and culture-negative groups. Mean age was 118.4 ± 55.3 months in culture-positive patients and 121.8 ± 55.9 months in culture-negative patients (p=0.770). Male sex was present in 59.2% and 51.1%, respectively (OR, 1.39; 95% CI, 0.61–3.14; p=0.432). The unadjusted comparison of demographic, clinical, and laboratory characteristics is summarized in Table 4. Culture-positive children more often had fever with arthralgia, anemia, elevated CRP, and elevated AST. Elevated AST showed the largest unadjusted effect estimate (OR, 6.00; 95% CI, 2.02–17.82), followed by anemia (OR, 5.45; 95% CI, 1.67–17.78) and elevated CRP (OR, 4.12; 95% CI, 1.46–11.58).

Table 4.

Demographic, Clinical, and Laboratory Variables According to Blood Culture Status

Variable	Culture Positive (n = 49)	Culture Negative (n = 45)	Unadjusted OR (95% CI)	p value
Age, months, mean ± SD	118.4 ± 55.3	121.8 ± 55.9	—	0.770
Male sex	29 (59.2)	23 (51.1)	1.39 (0.61–3.14)	0.432
Fever with arthralgia	13 (26.5)	4 (8.9)	3.70 (1.11–12.37)	0.026
Hepatosplenomegaly	8 (16.3)	5 (11.1)	1.56 (0.47–5.18)	0.464
Anemia	17 (34.7)	4 (8.9)	5.45 (1.67–17.78)	0.003
Thrombocytopenia	8 (16.3)	2 (4.4)	4.20 (0.84–20.93)	0.062
Elevated CRP	19 (38.8)	6 (13.3)	4.12 (1.46–11.58)	0.005
Elevated AST	21 (42.9)	5 (11.1)	6.00 (2.02–17.82)	<0.001

Notes: Age was compared using Student’s t test. Categorical variables were compared using Pearson’s chi-square test or Fisher’s exact test, as appropriate. ORs and Wald 95% CIs were calculated from the reported 2×2 counts.

Abbreviations: AST, aspartate aminotransferase; CI, confidence interval; CRP, C-reactive protein; OR, odds ratio; SD, standard deviation.

Exploratory Multivariable Analysis

In the exploratory multivariable logistic regression, all six candidate variables were entered simultaneously (Table 5), and all 94 participants had complete data for these variables. Fever with arthralgia remained associated with blood culture positivity (aOR, 8.56; 95% CI, 2.18–33.62; p=0.002), as did elevated CRP (aOR, 4.23; 95% CI, 1.22–14.65; p=0.023) and elevated AST (aOR, 6.18; 95% CI, 1.78–21.40; p=0.004). Anemia and thrombocytopenia were not independently associated with culture positivity. Continuous IgM ELISA was also not associated with culture positivity after adjustment (aOR per NTU, 1.03; 95% CI, 0.99–1.08; p=0.169). The overall model was statistically significant (likelihood-ratio χ²=34.5; df=6; p<0.001), with a Nagelkerke R² of 0.410. VIFs were 1.08 for IgM ELISA, 1.13 for elevated CRP, 1.31 for thrombocytopenia, 1.36 for anemia, 1.05 for elevated AST, and 1.14 for fever with arthralgia. With 49 culture-positive events and six predictors (approximately 8.2 events per predictor), overfitting and model instability remain possible; the model is therefore interpreted only as exploratory.

Table 5.

Exploratory Multivariable Logistic Regression for Blood Culture Positivity

Variable	Coefficient (β)	Adjusted OR	95% CI	p value
Fever with arthralgia	2.147	8.56	2.18–33.62	0.002
Elevated CRP	1.442	4.23	1.22–14.65	0.023
Elevated AST	1.821	6.18	1.78–21.40	0.004
Anemia	1.257	3.52	0.74–16.66	0.113
Thrombocytopenia	0.337	1.40	0.17–11.45	0.753
Brucella IgM ELISA value (per NTU)	0.030	1.03	0.99–1.08	0.169

Notes: All 94 participants had complete data for the six model variables. VIFs ranged from 1.05 to 1.36. With 49 culture-positive events and six predictors, the model had approximately 8.2 events per predictor. The model is exploratory, may be susceptible to overfitting/model instability, and was not internally or externally validated.

Abbreviations: AST, aspartate aminotransferase; CI, confidence interval; CRP, C-reactive protein; OR, odds ratio.

Discussion

This study extends the laboratory characterization of pediatric acute brucellosis by examining continuous IgM and IgG ELISA values together with qualitative serologic patterns according to contemporaneous blood culture status. Three findings are central. First, IgM was positive in nearly all children, whereas 20 (21.3%) had negative IgG and four (4.3%) had equivocal IgG results. Second, IgM magnitude did not differ significantly by culture status, whereas continuous IgG values were lower in culture-positive children; however, the categorical frequency of IgG negativity did not differ significantly after equivocal results were excluded. Third, fever with arthralgia, elevated CRP, and elevated AST remained associated with culture positivity in the exploratory multivariable model. These findings describe within-case associations and do not establish diagnostic performance, temporal relationships, or causality.

The observed class-specific serologic patterns are compatible with established Brucella immunobiology, in which IgM generally appears before IgG.⁴ However, the present cross-sectional analysis did not measure antibody changes over time. Although all children had newly diagnosed acute disease with reported symptoms lasting no longer than three weeks, exact symptom duration within this interval was not consistently available. Therefore, the observed IgM-positive/IgG-negative patterns should be regarded as descriptive findings rather than direct evidence of an earlier infection time point. A negative IgG result alone should not be used to exclude acute pediatric brucellosis when clinical and epidemiological findings are compatible, and isolated IgM positivity should likewise be interpreted with complementary clinical and laboratory evidence because IgM may be affected by cross-reactivity.^4,8

Continuous IgG values were lower among culture-positive children, but the mechanism underlying this association was not directly tested. Variation in the timing of antibody maturation relative to recovery of viable organisms is one biologically plausible explanation, but this remains a hypothesis because exact symptom duration and serial antibody measurements were unavailable. In addition, IgG was not included in the pre-specified exploratory multivariable model. The observed IgG difference is therefore an unadjusted association and should not be interpreted as an independent risk factor for bacteremia. This caution is reinforced by the categorical analysis: after equivocal IgG results were excluded, IgG negativity was numerically more frequent among culture-positive children but did not reach statistical significance.

The IgM findings reinforce the distinction between serologic recognition of infection and recovery of organisms from blood. IgM reflects the host immune response rather than the concentration of viable bacteria in the bloodstream, and its numerical magnitude would not necessarily be expected to track blood culture status linearly. Previous work has shown that ELISA can contribute to brucellosis diagnosis when IgM and IgG are interpreted together, but these studies generally address disease recognition rather than discrimination of culture status among already confirmed cases.^6,12,15,16 Similarly, an assay-specific IgM threshold reported to predict STAT positivity should not be extrapolated to blood culture positivity because the endpoints are biologically and analytically different.¹⁷ The present study did not identify or validate an IgM or IgG threshold for predicting bacteremia.

Blood culture was positive in 52.1% of the present cohort. Published pediatric series from Türkiye have reported lower culture-positive proportions, although recovery varies widely according to disease stage, specimen volume, incubation procedures, and antimicrobial exposure.^9–11 Culture remains particularly valuable because it provides etiologic confirmation and an isolate for further characterization, but a negative blood culture cannot exclude active brucellosis.^4,5 The present findings therefore support complementary rather than interchangeable roles for blood culture and class-specific serology.

The clinical and laboratory findings provide additional context. Elevated CRP and AST and the combination of fever with arthralgia remained associated with culture positivity, consistent with previous pediatric reports linking bacteremic brucellosis with a more pronounced inflammatory and hepatic phenotype.^9–11 These associations were exploratory and should not be used as a clinical prediction rule. The raw calendar-month distribution was retained only as a descriptive representation. Because observation time was unequal across calendar months, the figure cannot establish seasonal variation in pediatric brucellosis.

Strengths and Limitations

Strengths of this study include a well-characterized pediatric cohort from an endemic region, complete paired blood culture and serologic testing at the index presentation, and availability of numerical values for both IgM and IgG. Evaluating continuous antibody values alongside qualitative patterns preserved information that would have been lost by reducing ELISA results to positive/negative categories alone. The analysis also explicitly avoided treating culture-negative children as disease-negative controls.

Several limitations should be considered. The retrospective, single-center, cross-sectional design and modest sample size limit precision and generalizability. The findings derive from children in an endemic region of southeastern Türkiye and from isolates identified as B. melitensis; they may not apply directly to other Brucella species, non-endemic settings, very young pediatric subgroups, or patients with symptoms lasting longer than three weeks. Because all participants were already classified as having brucellosis and blood culture contributed to case ascertainment for some participants, the study cannot estimate ELISA sensitivity, specificity, predictive values, likelihood ratios, or diagnostic accuracy. Exact symptom duration within the three-week window and pre-referral antimicrobial exposure were not consistently available. Prior antimicrobial therapy could reduce blood culture recovery and may also influence observed serologic patterns. Patient-level blood volume inoculated into culture bottles, number of bottles/specimens, time from presentation to culture collection, and time-to-positivity were also not consistently available; these variables can materially affect Brucella recovery and should be systematically recorded in prospective studies. IgG analyses were secondary and were not incorporated post hoc into the pre-existing exploratory multivariable model; thus, the lower continuous IgG values observed among culture-positive children remain an unadjusted association. The six-predictor regression had approximately 8.2 culture-positive events per predictor and was not validated, so overfitting and model instability are possible despite low VIF values. STAT endpoint titration above 1:160 was not performed, preventing meaningful continuous correlation analyses between STAT titre, ELISA NTU values, and culture status. Molecular confirmation of culture isolates was unavailable.

Conclusion

Among children with confirmed acute brucellosis, IgM ELISA was positive in nearly all patients. Continuous IgG values were lower among culture-positive than culture-negative children, whereas IgM values did not differ significantly according to culture status; the categorical frequency of IgG negativity also did not differ significantly after equivocal results were excluded. These observations describe associations with contemporaneous blood culture status and should not be interpreted as evidence of antibody kinetics or as diagnostic-performance estimates. Absolute IgM or IgG ELISA values should not be used to predict blood culture positivity or to determine whether blood culture should be obtained in suspected pediatric brucellosis. Prospective studies incorporating precise symptom duration, prior antimicrobial exposure, standardized blood culture sampling, serial serologic measurements, and age-stratified analyses, particularly comparing younger children with older pediatric age groups, are needed to clarify these associations.

Funding Statement

The author received no financial support for the research, authorship, or publication of this article.

Data Sharing Statement

The deidentified data supporting the findings of this study are not publicly deposited because they contain sensitive pediatric clinical information and are subject to patient privacy and institutional ethical restrictions. Requests for access may be directed to the corresponding author at asuman.akar@dicle.edu.tr and will be considered subject to institutional ethics approval and, where required, a formal data-sharing agreement.

Ethics Approval and Informed Consent

The study was approved by the Institutional Review Board of Health Sciences University Gazi Yaşargil Training and Research Hospital (approval no. 165; September 13, 2024) and conducted in accordance with the Declaration of Helsinki. The requirement for written informed consent was waived by the Institutional Review Board because of the retrospective design and use of anonymized clinical data.

Author Contributions

The author made a significant contribution to the work reported, including the conception, study design, execution, acquisition of data, analysis and interpretation; took part in drafting, revising, and critically reviewing the article; gave final approval of the version to be published; agreed on the journal to which the article has been submitted; and agrees to be accountable for all aspects of the work.

Disclosure

The author reports no conflicts of interest in this work.