Seroprevalence of anti-ToxoplasmaIgG among the human population in Indonesia: a systematic review and meta-analysis

Abstract

Background

Toxoplasma gondii is a ubiquitous parasite that can cause significant complications when it infects pregnant women and immunocompromised patients. These complications include miscarriage, fetal abnormalities, and fatal cerebral toxoplasmosis. Despite its significance, the true burden of toxoplasmosis in Indonesia remains underexplored. Toxoplasmosis is usually diagnosed by detecting anti-Toxoplasma antibodies, especially IgG. Therefore, we aim to assess the seroprevalence of anti-Toxoplasma IgG among the human population in Indonesia. In addition, we assessed whether the seroprevalence differed across geographical regions, populations, or population risk levels. Its correlation with annual precipitation was also assessed.

Methods

Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we conducted a systematic review and meta-analysis of data retrieved from PubMed, Portal Garuda, Neliti, and Onesearch.id. Additionally, Google Scholar, government repositories, and the reference list of studies were searched for additional data. We pooled seroprevalence data using the inverse-variance method and a random effects model. Heterogeneity was assessed using I² statistics and Cochran’s Q test. Risk-of-Bias (RoB) was evaluated using the Joanna Briggs Institute Checklist for Prevalence Studies. Publication bias was assessed using Doi plots and the Luis Furuya-Kanamori (LFK) index. We performed subgroup analysis, meta-regression, and sensitivity analysis to explore source heterogeneity and the robustness of the pooled estimates. We used Spearman's rank correlation coefficient to assess the correlation between seroprevalence and annual precipitation.

Result

In total, 56 studies were included in this study. The adjusted seroprevalence of anti-Toxoplasma IgG was 60.06% (95% CI: 52.22–67.65%). Study location and detection method were detected as significant sources of heterogeneity by subgroup analysis but not meta-regression. However, subgroup analysis and meta-regression identified the study population and population risk level as significant sources of heterogeneity. Publication year, sample size, and RoB were identified as non-significant moderators. Seroprevalence did not correlate with annual precipitation.

Conclusion

Toxoplasmosis is highly prevalent among the human population in Indonesia; however, our study mainly relied on studies with small sample sizes. Furthermore, most of the studies were performed in Java; therefore, some high-quality population-based studies must be conducted in other regions of Indonesia to better estimate the prevalence of toxoplasmosis across the country.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-21317-2.

Background

Toxoplasma gondii is an apicomplexan parasite that infects most warm-blooded animals, including humans. Humans primarily become infected by consuming water, food, or soil contaminated with Toxoplasma oocyst and eating undercooked meats infected with Toxoplasma bradyzoite cyst [1]. In fact, the consumption of several meat-producing animals, such as pigs, cattle, goats, and poultry, has been implicated in substantially increasing the risk of toxoplasmosis in humans [2–6].

In otherwise healthy adults, Toxoplasma infection, known as toxoplasmosis, typically results in asymptomatic chronic infection. However, in several populations, toxoplasmosis may lead to debilitating or even fatal illness. For example, acute toxoplasmosis in pregnant women can lead to miscarriage, preterm birth, low birth weight, and fetal abnormalities [7, 8]. In patients with HIV, acute or reactivated toxoplasmosis may result in cerebral or disseminated toxoplasmosis, with cerebral toxoplasmosis having an estimated mortality rate of approximately 30% [9–11]. Survivors of cerebral toxoplasmosis may face persistent neurological deficits [12].

Generally, toxoplasmosis in humans is diagnosed by detecting the presence of anti-Toxoplasma antibodies such as IgM and IgG. IgM antibodies can be detected as early as one week after infection and may remain detectable for approximately two years [13, 14]. On the other hand, IgG antibodies appear one to three weeks after IgM becomes detectable and remain for life [14]. However, detecting anti-Toxoplasma IgM alone is insufficient to establish the diagnosis due to its lack of specificity [15, 16]. Therefore, anti-Toxoplasma IgG detection is required to confirm the diagnosis, even in cases where acute toxoplasmosis is suspected. The Sabin-Feldman dye test is considered the gold standard for anti-Toxoplasma antibody detection [17]. However, DT is impractical. As such, other diagnostic methods have been employed to detect anti-T. gondii antibody in both clinical and research settings. These include Enzyme-linked immunosorbent assay (ELISA), Latex Agglutination Test (LAT), and Lateral-flow Immunoassay (LFA), among others.

Around two billion people are infected by this parasite globally [18]. However, the seroprevalence of toxoplasmosis in humans varies between regions and communities, depending on social, economic, and cultural factors [19, 20]. For example, Jones et al. reported an adjusted seroprevalence of 22.5% among people over 12 years old in the United States [21]. A study in Germany estimated that 55% of adults in the country are seropositive for Toxoplasma [22]. However, much lower Toxoplasma seroprevalence was observed in China, with only 12.3% of the examined samples being positive [23].

A recent meta-analysis [24] estimated that the seroprevalence of Toxoplasma infection in Indonesia is at least 60%, but this estimate was based on a single study [25]; the authors’ search strategy did not include Indonesian journal databases such as Portal Garuda, in which several articles on this subject are indexed. To address this gap, we conducted a systematic review and meta-analysis to assess the seroprevalence of anti-Toxoplasma IgG in the human population in Indonesia. In addition, we assessed whether the seroprevalence differs across geographical regions, populations, or population risk levels.

Methods

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [26] and is registered with PROSPERO (ID CRD42023466909). The PRISMA 2020 checklist is available in Additional File 1.

Eligibility criteria

We included studies that met the following criteria:

1. Reported the seroprevalence of Anti-Toxoplasma IgG in the human population in Indonesia or presented data that can be used to estimate the seroprevalence

2. Detailed the diagnostic methods used to estimate the seroprevalence of Anti-Toxoplasma IgG in the human population in Indonesia

3. Were published before November 2024

We excluded studies that met the following criteria:

1. Did not contain primary data (e.g., systematic review and meta-analysis).

2. Were case reports.

3. Had limited access and failed to respond to follow-up emails twice.

Search strategy

We systematically searched PubMed, Portal Garuda, Neliti, and Onesearch.id for eligible records. Key search terms used in different combinations were toxoplasmosis, Toxoplasma, prevalence, epidemiology, seroprevalence, seroepidemiology, and Indonesia. Boolean operators and truncation were used to increase the sensitivity of the search strategy. No language or study type restrictions were applied. All identified records were retrieved and uploaded to Rayyan [27] for review.

Additionally, we did a manual search in the Indonesian Ministry of Health Repository, the National Research and Innovation Agency (BRIN) Repository, The National Institute of Health Research and Development (Balitbangkes) Repository, and Google Scholar for eligible records. The reference list of studies eligible for further assessment and all review articles were also screened for potentially relevant studies.

A detailed search strategy can be found in Additional File 2.

Selection process

After deduplication, two reviewers (SK and AHD) independently selected records that satisfied the eligibility criteria. Any disagreements between the reviewers were resolved through discussion. Afterwards, a third reviewer (TMP) confirmed the eligibility of the records included. The reviewers were not blinded to the journal titles, study authors, or institutions.

Data extraction

In this review, Toxoplasma infection is defined as the detection of anti-Toxoplasma IgG through serological testing of the human population. The primary outcome of this review is the pooled seroprevalence of Toxoplasma infection in the human population in Indonesia through a systematic review. Seroprevalence is the ratio of positive samples to the total number of samples. The additional outcomes of this study were the seroprevalence of Toxoplasma infection across different geographical regions, study populations, and population risk levels. The study locations were grouped based on the major geographical regions they belong to: Java, Sumatra, Kalimantan, Sulawesi and Maluku, Lesser Sunda Island, and Papua. The population risk level was stratified into high- and low-risk groups [28]. Populations that belonged to the high-risk group included women with a history of miscarriage and/or stillbirth, people who routinely have close contact with animals (e.g., animal market workers, pet shop workers), cat owners, immunocompromised patients, and psychiatric patients. Populations that did not belong to these groups were classified as low risk. Additionally, we also assessed the correlation between the toxoplasmosis seroprevalence and annual precipitation at national and local level. The national annual precipitation data were retrieved from ourworldindata website on December 24, 2024 (publicly available from https://ourworldindata.org/grapher/average-precipitation-per-year) [29]. Meanwhile, the local annual precipitation data were retrieved from publicly available data on the local Central Bureau of Statistics (BPS) website.

We extracted the following data from the studies: first author’s last name, year of publication, study location, population studied, diagnostic method, sample size, and total number of cases. If the number of cases was not presented directly, we inferred it based on the reported prevalence. The two reviewers (SK and AHD) extracted these data into a Google Sheets spreadsheet. A third author (TMP) referred to the original studies to confirm whether these data had been extracted correctly.

Quality assessment

The quality of the records that were included was assessed independently by two authors (AMTS and FPSW), using a checklist for prevalence studies developed by the Joanna Briggs Institute [30]. Any disagreements between the reviewers were resolved through discussion. When assessing the question, “Was the sample size adequate?”, we estimated the minimum sample size via the formula [n = z^2 * (p * (1-p))/e^2], where z was set at 95%, e was set at 5%, and p was set at 16.4% following the estimated prevalence of toxoplasmosis in Asia [24]. Based on these findings, we set the minimum sample size at 211 participants and used this value as our evaluation standard. A score was calculated based on the proportion of “yes” answers. Afterwards, the records were classified as “low risk of bias,” “moderate risk of bias,” or “high risk of bias” if their scores were ≥ 70%, 50–69%, or ≤ 49%, respectively [31, 32]. Studies with a high risk of bias (RoB) were excluded from further analysis.

Data analysis

Narrative and quantitative approaches were used to synthesize systematic review data. Unless otherwise stated, all data processing and analysis were conducted via Meta (version 7.0–0) and MetaSens (version 1.5–2) packages in R computational package (version 4.4.2) [33]. The prevalence estimate was calculated for each study, and we then pooled the seroprevalence data via the inverse-variance method and a random effects model. The summary measure was calculated using the Freeman-Tukey Double arcsine transformation. We also calculated the 95% confidence interval (95% CI). A summary of the results and the heterogeneity among the studies was presented via a forest plot. I² statistics were used to determine heterogeneity, with a value of > 75% considered substantial heterogeneity [32]. Cochran’s Q test was used to test the significance of heterogeneity. Subgroup analysis and meta-regression were performed to explore potential sources of heterogeneity. In addition, sensitivity analysis was performed to assess any study that might impact the pooled prevalence estimate. Doi plots and the Luis Furuya-Kanamori (LFK) index were used to assess the presence of publication bias [34]. A p-value was considered significant if it was < 0.05.

The analysis on the correlation between toxoplasmosis seroprevalence and annual precipitation was performed using native functions in R computational package (version 4.4.2). First the distribution of seroprevalence data was tested using Shapiro–Wilk test. Next, Spearman's rank correlation coefficient was calculated to determine the correlation of toxoplasmosis seroprevalence and annual precipitation. Finally, a scatter plot with trend line was produced using ggplot2 (version 3.5.1) and smplot2 (version 0.2.4) packages in R computational package (version 4.4.2) [35, 36].

Results

Our database search identified 612 records (Fig. 1), 102 of which passed the initial screening and were retrieved for further assessment. However, we were unable to retrieve 8 of these records. Among these, 46 were included in the subsequent analysis. Additionally, we identified 4609 records through manual search in government repositories, Google Scholar search, and manual citation searching. Through this method, we found ten records that met our inclusion criteria and were included in the downstream analysis. Taken together, 56 studies were included in this study (Table 1).

Fig. 1.

PRISMA flow diagram. Of the 5,221 identified records, 56 were included

Table 1.

List and characteristics of studies that assessed the anti-Toxoplasma IgG seroprevalence in Indonesia

Year	Author	Location	Region	Population	Method	Case	Sample	Seroprevalence	ROB	Ref
1976	Durfee	South Kalimantan Province	Kalimantan	General Population	IHA	330	1050	31.43%	Low	[37]
1996	Uga	Sidoarjo Regency	Java	General Population	LAT	156	244	63.93%	Low	[38]
2000	Konishi	Surabaya City	Java	General Population	ELISA	1021	1761	58.00%	Medium	[39]
2002	Sardjono	Malang City	Java	Women with History of Miscarriage and/or Stillbirth	ELISA	22	43	51.16%	Low	[40]
				Pregnant Women	ELISA	14	23	60.87%
2003	Terazawa	Jakarta	Java	General Population	ELISA	1185	1693	69.99%	Low	[25]
2004	Murkati	Surakarta City	Java	Women with History of Miscarriage and/or Stillbirth	ELISA	33	100	33.00%	Medium	[41]
2009	Sanjaya	Manado City	Sulawesi and Maluku	General Population	LAT	22	50	44.00%	Medium	[42]
2011	Lestari	National	National	General Population	ELISA	2036	3067	66.40%	Low	[43]
2012	Tolistiawaty	Palu City	Sulawesi and Maluku	General Population	ELISA	177	412	42.96%	Low	[44]
2013	Laksemi	Denpasar City	Lesser Sunda Islands	General Population	ELISA	391	790	49.49%	Low	[2]
2013	Wiyarno	Surabaya City	Java	People with Close Contact with Animals	ELISA	16	20	80.00%	Medium	[45]
				General Population		9	20	45.00%
2014	Prasetyo	Surakarta City	Java	General Population	ELISA	39	130	30.00%	Medium	[46]
2014	Resnhaleksmana	West Nusa Tenggara Province	Lesser Sunda Islands	Psychiatric Patients	ELISA	23	42	54.76%	Medium	[47]
2014	Sari	Bantul Regency	Java	Cat Owner	ELISA	7	11	63.64%	Medium	[48]
				General Population		43	79	54.43%
2015	Agustin	Surabaya City	Java	Cat Owner	ELISA	13	25	52.00%	Low	[49]
				General Population		12	25	48.00%
2015	Krihariyani	Surabaya City	Java	Pregnant Women	ELISA	26	40	65.00%	High	[50]
2015	Prasetyo	Central Java	Java	HIV Patients	ELISA	260	597	43.55%	Low	[51]
2015	Raharjo	Surakarta City	Java	HIV Patients	ELISA	9	51	17.65%	Low	[52]
2015	Sari	Central Java	Java	General Population	ELISA	150	357	42.02%	Low	[53]
2016	Dwinata	Badung Regency	Lesser Sunda Islands	Pregnant Women	ELISA	36	330	10.91%	Medium	[54]
2016	Kurniawati	Surabaya City	Java	General Population	ELISA	85	234	36.32%	Low	[55]
2016	Laksmi	Karangasem Regency	Lesser Sunda Islands	General Population	ELISA	14	106	13.21%	Medium	[56]
2016	Lestari	Minahasa Regency	Sulawesi and Maluku	General Population	LAT	15	22	68.18%	High	[57]
2016	Seran	Minahasa Regency	Sulawesi and Maluku	General Population	LAT	11	22	50.00%	Low	[58]
2016	Tuda	Manado City	Sulawesi and Maluku	General Population	LAT	72	135	53.33%	Medium	[59]
				Pregnant Women		47	109	43.12%
2017	Amalia	Surabaya City	Java	General Population	ELISA	14	125	11.20%	Medium	[60]
2017	Febianingsih	Gianyar Regency	Lesser Sunda Islands	General Population	ELISA	136	240	56.67%	Low	[61]
2017	Iskandar	Malang City	Java	General Population	ELISA	50	57	87.72%	High	[62]
2017	Retmanasari	Central Java	Java	General Population	ELISA	394	630	62.54%	Low	[20]
2017	Tuda	North Sulawesi, Gorontalo, and North Maluku Province	Sulawesi and Maluku	General Population	LAT	501	856	58.53%	Low	[63]
2018	Muflikhah	Sleman Regency	Java	Psychiatric Patients	ELISA	65	94	69.15%	Medium	[3]
				General Population		42	64	65.63%
2019	Alviyah	Surabaya City	Java	General Population	ELISA	116	132	87.88%	Medium	[64]
2019	Arwie	Makassar City	Sulawesi and Maluku	People with Close Contact with Animals	LAT	4	10	40.00%	High	[65]
2019	Darmawan	Jambi City	Sumatra	General Population	LFA	17	41	41.46%	Medium	[66]
2019	Halleyantoro	Jakarta	Java	HIV Patients	ELISA	34	88	38.64%	Medium	[67]
2020	Dzikriana	Surabaya City	Java	People with Close Contact with Animals	LFA	26	30	86.67%	Medium	[68]
2020	Fihiruddin	Sleman Regency	Java	General Population	ELISA	224	385	58.18%	High	[69]
2020	Humaryanto	Jambi City	Sumatra	General Population	LAT	36	60	60.00%	High	[70]
2020	Marthalia	Surabaya City	Java	Cat Owner	ELFA	11	19	57.89%	Medium	[71]
2020	Utami	Batu City	Java	People with Close Contact with Animals	MAT	40	45	88.89%	Low	[72]
2021	Harianja	Samarinda City	Kalimantan	Cat Owner	ECLIA	10	23	43.48%	High	[73]
2021	Polanunu	Makassar City	Sulawesi and Maluku	Pregnant Women	ELISA	60	184	32.61%	Low	[74]
2023	Afrianti	Semarang City	Java	Cat Owner	ELFA	6	25	24.00%	High	[75]
				General Population		2	25	8.00%
2023	Hanina	Jambi City	Sumatra	Women with History of Miscarriage and/or Stillbirth	ELFA	22	52	42.31%	Medium	[76]
2023	Jayawardhana	Sidoarjo Regency	Java	HIV patients	MAT	4	30	13.33%	High	[77]
2023	Meirawan	Jakarta	Java	HIV patients	ELISA	38	56	67.86%	Medium	[78]
2023	Novikasari	Sampang Regency	Java	Pregnant Women	LFA	5	25	20.00%	Medium	[79]
2023	Pratama	South Tangerang City	Java	People with Close Contact with Animals	LFA	11	25	44.00%	Medium	[80]
2023	Riansari	Semarang City	Java	General Population	ELISA	43	88	48.86%	Low	[81]
2023	Sari	Denpasar City	Lesser Sunda Islands	Pregnant Women	ELFA	9	44	20.45%	Medium	[82]
2023	Setia	Malang City	Java	HIV Patients	ELISA	87	87	10.,00%	Medium	[83]
2024	Afrianti	Semarang City	Java	General Population	ECLIA	47	177	26.55%	Medium	[84]
2024	Rahman	Yogyakarta City	Java	General Population	LFA	14	25	56.00%	High	[85]
2024	Widyaswara	Magelang Regency	Java	General Population	LFA	0	25	0.00%	High	[86]
2024	Wikandari	Semarang City	Java	General Population	ELISA	31	87	35.63%	Medium	[87]
2024	Wulandari	Yogyakarta City	Java	General Population	LFA	13	69	18.84%	High	[88]

IHA Indirect Hemagglutination Test, ELFA Enzyme-linked fluorescence assay, ELISA Enzyme-linked immunosorbent assay, LAT Latex Agglutination Test, MAT Modified Agglutination Test, LFA Lateral-flow Immunoassay, ECLIA Enhanced chemiluminescence immunoassay, RoB Risk-of-Bias

Quality assessment

The RoB for the included studies was assessed via a checklist for prevalence studies developed by the Joanna Briggs Institute [30]. Overall, 19 studies (33.93%) had a low RoB, 25 studies (44.64%) had a moderate RoB, and 12 studies (21.43%) had a high RoB (Additional file 3). Most studies (n = 40) had sample sizes of fewer than 211 participants. Studies that were high RoB, were excluded from subsequent analyses.

Pooled seroprevalence

Pooled seroprevalence was derived from the data of 14,715 participants [2, 3, 20, 25, 37–39, 41–49, 52–56, 58–61, 63, 64, 66–68, 71, 72, 74, 76, 78–84, 87]. Among them, 7982 cases were identified. According to the random effects model (Fig. 2), the pooled seroprevalence of toxoplasmosis in Indonesia was 48.58% (95%CI: 41.52%–55.67%). Based on the prediction interval, future studies are expected to yield a seroprevalence rate of between 8 and 90%. Additionally, the pooling result was substantially heterogeneous (I² = 97.6% (95% CI: 96.1%–97.1%); p < 0.01). Subgroup analysis and meta-regression were performed to investigate the source of heterogeneity.

Fig. 2.

Forest plot depicting the pooled seroprevalence of toxoplasmosis in Indonesia

Subgroup analysis

Subgroup analysis was performed based on the region where the study was conducted, the study population, the population risk level, the sample size, the detection method, and the RoB. Some of the studies derived their data from different populations. Therefore, the number of data sources for the subgroup analysis differed from the number of studies included.

By region

In total, 28 studies were performed in Java, two in Sumatera, one in Kalimantan, six in the Lesser Sunda Islands, and six in Sulawesi and Maluku (Table 2; Fig. 3; Additional file 4). No study that can be included in the subgroup analysis has been performed in Papua. The pooled seroprevalence rates of toxoplasmosis in Java, Sumatera, Kalimantan, Lesser Sunda Island, and Sulawesi and Maluku were 53.15% (95% CI: 43.47–62.72%), 41.93% (95% CI: 31.97%—52.22%), 31.43% (95%CI: 28.65%–34.27%), 32.56% (95% CI: 16.10%–51.56%), and 46.08% (95% CI: 38.04%–54.22%), respectively. Significant differences between subgroups were observed (p < 0.0001).

Table 2.

Subgroup analysis results

Subgroup Name	Number of data source	Pooled Prevalence	95%CI	I2	Subgroup Differences
By Region
Java	28	53.15%	43.47%–62.72%	96.9%	P < 0.0001
Sumatra	2	41.93%	31.97%–52.22%	0.0%
Kalimantan	1	31.43%	28.65%–34.27%	-
Lesser Sunda Islands	6	32.56%	16.10%–51.56%	98.1%
Sulawesi and Maluku	6	46.08%	38.04%–54.22%	91.2%
By Study Population
General Population	27	47.85%	40.72%–55.01%	97.6%	P = 0. 0026
Cat Owner	3	56.40%	42.62%–69.73%	0.0%
People with routine close contacts with animals	4	76.88%	55.51%–93.10%	82.7%
Pregnant women	6	29.39%	16.22%–44.51%	93.6%
Women with history of miscarriage and/or stillbirth	3	40.85%	30.49%–51.63%	53.7%
Immunocompromised patients	5	57.89%	23.06%–88.83%	98.3%
Psychiatric patients	2	63.21%	48.80%–76.55%	60.9%
By Population Risk Level
Low risk	33	44.57%	37.78%–51.46%	97.9%	P = 0.0392
High risk	17	59.99%	47.01%–72.33%	95.0%
By Sample Size
< 211 Samples	29	48.30%	38.31%–58.36%	95.9%	P = 0.9008
≥ 211 Samples	15	49.16%	40.72%–57.62%	98.7%
By Detection Method
IHA	1	31.43%	28.65%–34.27%	-	P < 0.0001
LAT	5	54.71%	47.22%–62.10%	75.2%
ELISA	29	48.64%	39.57%–57.75%	98.0%
LFA	4	48.71%	20.78%–77.05%	90.1%
MAT	1	88.89%	77.78%–96.71%	-
ELFA	3	38.39%	18.69%–60.16%	79.0%
ECLIA	1	26.55%	20.29%–33.33%	-
By ROB Assessment Result
Low risk of bias	19	50.98%	43.55%–58.39%	97.8%	P = 0.5356
Moderate risk of bias	25	46.71%	35.64%–57.94%	97.4%

IHA Indirect Hemagglutination Test, ELISA Enzyme-linked immunosorbent assay, LAT Latex Agglutination Test, LFA Lateral-flow Immunoassay, MAT Modified Agglutination Test, ELFA Enzyme-linked fluorescence assay, ECLIA Enhanced chemiluminescence immunoassay, RoB Risk-of-Bias

Fig. 3.

Geographical map of seroprevalence rates of toxoplasmosis by regions

One report derived its data from a nationwide study conducted from 2007 to 2008 [43]. While this report described toxoplasmosis seroprevalence by province, the data it provided could not be pooled in the subgroup analysis. Based on their data, toxoplasmosis seroprevalence in Java, Sumatera, Kalimantan, Lesser Sunda Island, Sulawesi and Maluku, and Papua ranges between 66.2%–75.5%, 61.4%–86%, 60.2%–71.3%, 30.5%–65.0%, 57.7%–75.6%, and 50.0%–91.7%, respectively.

By study population

The data from the studies were derived from seven groups of study populations (Table 2; Additional file 4). These include the general population (n = 27), cat owners (n = 3), people who routinely had close contact with animals (n = 4), pregnant women (n = 6), women with a history of miscarriage and/or stillbirth (n = 3), immunocompromised patients (n = 5), and psychiatric patients (n = 2). Unintentionally, the immunocompromised group comprised only HIV-infected individuals. The pooled prevalence rates for each of these population groups were 47.85% (95% CI: 40.72%–55.01%) for the general population, 56.40% (95% CI: 42.62%–69.73%) for cat owners, 76.88% (95% CI: 55.51%–93.10%) for people who routinely had close contact with animals, 29.39% (95% CI: 16.22%–44.51%) for pregnant women, 40.85% (95% CI: 30.49%–51.63%) for women with a history of miscarriage and/or stillbirth, 57.89% (95% CI: 23.06%–88.83%) for immunocompromised patients, and 63.21% (95% CI: 48.80%–76.55%) for psychiatric patients. Significant differences between the subgroups were observed (p = 0. 0026).

By population risk level

We found 17 data points from the high-risk group and 33 from the low-risk group (Table 2; Additional file 4). The pooled seroprevalence of toxoplasmosis in the high-risk group was 59.99% (95% CI: 47.01%–72.33%), whereas that in the low-risk group was 44.15% (95% CI: 37.78%–51.46%). We observed no differences between the subgroups (p = 0. 0392).

By sample size

We separated the studies by whether they passed our minimum sample size (211 samples) or not. As a result, 15 studies passed the threshold, whereas 29 did not (Table 2; Additional file 4). The pooled seroprevalence of studies that passed the threshold was 49.16% (95% CI: 40.72%–57.62%), whereas studies that did not pass the threshold had a seroprevalence of 48.30% (95% CI: 38.31%–58.36%). We observed no statistically significant differences between the subgroups (p = 0. 9008).

By detection method

The studies assessed the seroprevalence of toxoplasmosis using an indirect hemagglutination test (IHA; one study), a latex agglutination test (LAT; five studies), enzyme-linked immunosorbent assay (ELISA; 29 studies), a lateral flow assay (LFA; five studies), a modified agglutination test (MAT; one studies), an enzyme-linked fluorescence assay (ELFA; three studies), and an enhanced chemiluminescence immunoassay (ECLIA; one studies). The pooled seroprevalence rates based on IHA, LAT, ELISA, LFA, MAT, ELFA, and ECLIA were 31.43% (95% CI: 28.65%–34.27%), 54.71% (95% CI: 47.22%–62.10%), 48.64% (95% CI: 39.57%–57.75%), 48.71% (95% CI: 20.78%–77.05%), 88.89% (95% CI: 77.78%–96.71%), 38.39% (95% CI: 18.69%–60.16%), and 26.55% (95% CI: 20.29%–33.33%), respectively (Table 2; Additional file 4). Substantial differences between the subgroups were observed (p < 0.0001).

By RoB

Most of the studies had a medium RoB (25), followed by a low RoB (19). The pooled seroprevalence rate based on the RoB assessment results was 50.98% (95% CI: 43.55%–58.39%) for studies with a low RoB and 46.71% (95% CI: 35.64%–57.94%) for studies with a medium RoB (Table 2; Additional file 4). No significant differences between the subgroups were observed (p = 0. 5356).

Meta-regression

To further explore potential sources of heterogeneity, we performed meta-regression. From this analysis, publication year (R² = 0.00%; QM(df = 1) = 0.0581; p = 0.8096) and sample size (R² = 0.00%; QM(df = 1) = 0.9339; p = 0.3339) were identified as non-significant moderators (Additional file 5).

Additionally, we performed meta-regression on subgroups with significant subgroup differences: region, study population, population risk level, and detection methods. Study population (R² = 14.19%; QM(df = 6) = 13.5075; p = 0.0356) and population risk level (R² = 7.54%; QM(df = 1) = 5.0388; p = 0.0248) were identified as significant moderators. Meanwhile, region (R² = 1.66%; QM(df = 1) = 4.6536; p = 0.3247) and detection method (R² = 0.00%; QM(df = 6) = 5.5461; p = 0.4759) were identified as non-significant moderators.

Sensitivity analysis

To determine the robustness of the pooling result, we performed a sensitivity analysis using three scenarios and a leave-one-out analysis. In the first scenario, we included eight studies [4, 89–95] that assessed the seroprevalence of toxoplasmosis but did not specify their detection methods. Therefore, these studies did not meet our inclusion criteria. In the second scenario, we included all studies, regardless of their ROB, in the meta-analysis. In the third scenario, we redid the pooling by excluding five studies that had less than 40 samples. The results of the three scenarios were 46.75% (95% CI: 40.36%–53.19%), 47.30% (95% CI: 40.65%–54.00%), and 48.58% (95% CI: 41.52%–55.67%), respectively. Furthermore, when we performed a leave-one-out analysis, the meta-analysis results remained stable after re-evaluation. Detailed sensitivity analysis results can be seen in Additional file 6.

Publication bias analysis

We analyzed publication bias on studies that used serological tests with the LFK index and Doi plot. We detected publication bias in favor of studies with lower seroprevalence rates (LFK index = − 1.83; Fig. 4). We then performed a trim-and-fill analysis, which assigned 13 studies (Fig. 5). This resulted in a pooled seroprevalence of 60.06% (95% CI: 52.22–67.65%; τ²: 0.0873; I²: 98.4%).

Fig. 4.

A Doi plot examining the risk of publication bias for the seroprevalence of toxoplasmosis. Note the left-side asymmetry as indicated by the LFK index

Fig. 5.

A Doi plot examining the risk of publication bias for the seroprevalence of toxoplasmosis corrected using the trim-and-fill method

Correlation between toxoplasmosis seroprevalence and annual precipitation

We examined toxoplasmosis seroprevalence from 42 studies for its correlation with national-level annual precipitation data [29]. On the other hand, due to the limited availability of local-level annual precipitation data [96–112], the correlation of local-level annual precipitation data with toxoplasmosis seroprevalence was examined in only 26 studies. Both national (R = 0.05, p = 0.753) and local (R = 0.065, p = 0.752) annual precipitation were not correlated with toxoplasmosis seroprevalence (Fig. 6). See Additional File 7 for the detailed annual precipitation data.

Fig. 6.

Scatterplots with trend line depicting the correlation between toxoplasmosis seroprevalence and national (A) and local (B) annual precipitation data

Discussion

Based on the available literature, this systematic review and meta-analysis were conducted to approximate the true seroprevalence of toxoplasmosis in Indonesia. Unlike other systematic reviews on similar topics, we included Portal Garuda, Neliti, and Onesearch.id, databases that index literature published in Indonesian journals and repositories. Additionally, we also searched government reports published in several repositories managed by Indonesian government bodies. Given that many of these studies were published in the Indonesian language, our attempt allowed us to minimize the occurrence of language bias in our work. In this study, we estimated the seroprevalence of anti-Toxoplasma IgG among humans in Indonesia based on data from 14,715 participants with 7,982 cases. This yields a seroprevalence of 48.58%. However, 13 hypothetical studies were added to adjust for publication bias, yielding an adjusted seroprevalence of 60.06%. Based on our prediction interval, subsequent studies on the topic will yield an anti-Toxoplasma IgG seroprevalence of between 8 and 90%. This result is similar to another systematic review that estimated that the seroprevalence of toxoplasmosis in Indonesia was at least 60% [24]. Additionally, our finding aligns with an Indonesian nationwide biomedical survey in 2007 – 2008, which stated that national toxoplasmosis seroprevalence was 66.40% [43]. Our result was also comparable to other systematic reviews conducted in Thailand and Malaysia, which reported seroprevalence rates of between 2.6% and 53.7% in Thailand and between 10.6% and 59.7% in Malaysia [113, 114]. However, other neighboring countries, namely Cambodia, Singapore, Myanmar, Laos, and Vietnam, had much lower estimated seroprevalence rates of between 5.8% and 31.7% [114]. In addition, our estimated seroprevalence was higher than the overall estimated seroprevalence in Asia, which ranges from 16.4% to 29% [24, 115, 116].

The geographical region was revealed to significantly contribute to the heterogeneity of the meta-analysis result by subgroup analysis but not meta-regression. This means that the subgroup analysis detected significant variance in toxoplasmosis prevalence across different geographical regions. However, an examination of the national biomedical survey report [43] reveals that toxoplasmosis seroprevalence across these regions is, in fact, comparable with a slightly smaller seroprevalence range in Lesser Sunda Island. Therefore, the subgroup analysis may have falsely detected a difference in toxoplasmosis seroprevalence between the regions, considering that the chance of producing a false positive result increases with the number of independent subgroups [117]. On the other hand, the opposite might also be true, as most of the studies were conducted in Java. At the same time, the other regions have at most six studies. This would make meta-regression too underpowered to detect statistically significant associations [118]. Regardless, given Indonesia's vast geographical size and extensive sociocultural diversity, it is plausible that there are actual variations in toxoplasmosis seroprevalence across the region. Different parts of Indonesia are known to have variations in living standards, cultural practices, and precipitation patterns, among others. These factors have been documented to influence the risk of toxoplasmosis both in Indonesia and other countries[56, 58, 119]. As such, it is important to investigate by conducting high-quality population-based studies outside Java to obtain a more accurate assessment of the true prevalence of toxoplasmosis in these regions.

Another factor that might affect the difference in toxoplasmosis seroprevalence between the study regions is difference in annual precipitation. Previous studies have associated the difference in annual precipitation with the incidence of congenital toxoplasmosis and reactivation of toxoplasmic retinochoroiditis [120, 121]. This is likely related to the increased survival of T. gondii oocyst in humid soil [122]. However, our analysis suggested that toxoplasmosis seroprevalence was not related to difference in annual precipitation on both national and local level. This is likely because we assessed the seroprevalence of chronic toxoplasmosis as measured by IgG seropositivity. Meanwhile, the other studies [120, 121] assessed active and reactivated toxoplasmosis. Therefore, we hypothesize that while the difference in annual precipitation may be a risk factor for congenital toxoplasmosis and reactivation of toxoplasmic retinochoroiditis, it is not a significant risk factor for chronic toxoplasmosis. Regardless, further studies assessing the association between active toxoplasmosis in Indonesia and annual precipitation is needed to clarify this finding.

Studies being analyzed in this systematic review used different diagnostic tools to detect anti-Toxoplasma IgG. These diagnostic tools were IHA, LAT, ELISA, LFA, MAT, ELFA, and ECLIA. The difference between the diagnostic tools being employed by these studies was detected to be a significant moderator by subgroup analysis but not meta-regression. One of the simplest explanations would be the difference in their sensitivity and specificity. However, several studies found that these different diagnostic tools have comparable sensitivity and specificity to detect anti-Toxoplasma IgG [13, 123]. As such, another possible explanation would be that the subgroup analysis falsely detected differences across studies using these different diagnostic tools. This suspicion is reinforced by the subgroup analysis, which included seven independent subgroups, substantially increasing the likelihood of a false positive result [117].

Study population and population risk level are two moderators that significantly contributed to the heterogeneity as assessed by subgroup analysis and meta-regression. These moderators accounted for 14.19% and 7.54% of the heterogeneity. In our study, high-risk populations comprised women with a history of miscarriage and/or stillbirth, people who routinely have close contact with animals (e.g., animal market workers, pet shop workers), cat owners, immunocompromised patients, and psychiatric patients. These population groups are people with a higher probability of contracting toxoplasmosis [28]. In line with this assumption, we found that the high-risk group had higher anti-Toxoplasma IgG seroprevalence.

We estimated that the seroprevalence of toxoplasmosis among women with a history of miscarriage and/or stillbirth was 40.85%. These data were derived from three studies that reported 195 cases in total. Our findings are similar to several other meta-analyses that investigated similar conditions. The reported seroprevalence in this population ranged from 32 to 43% [28, 124, 125]. Indeed, toxoplasmosis has long been reported as one of the risk factors for miscarriage, with IgM seropositivity resulting in a higher chance of miscarriage than IgG seropositivity alone [7, 125]. However, our study did not assess active toxoplasmosis, which is indicated by IgM.

Cats are the definitive hosts of Toxoplasma. As such, they are believed to be a significant risk factor for toxoplasmosis through close contact with them [126, 127]. In addition, given that Toxoplasma can infect virtually all warm-blooded animals, people with occupations that put them in close contact with animals are also thought to be at increased risk of contracting the disease [126]. In our study, the estimated seroprevalence of toxoplasmosis among cat owners and people who routinely have close contact with animals was higher than the general population (56.40%, 76.88%, and 47.85%, respectively). However, it should be noted that the studies that assess the toxoplasmosis seroprevalence, specifically among cat owners and people who routinely have close contact with animals in the current review, only included a small number of participants. This potentially makes them unable to assess the true prevalence accurately in the population. As such, extrapolating this finding must be performed carefully. Regardless, a large population-based study assessing this issue might be needed to estimate the seroprevalence of toxoplasmosis among cat owners in Indonesia more accurately.

Our study revealed an alarmingly high seroprevalence of toxoplasmosis among HIV-infected individuals (57.89%). This prevalence is much higher than the estimated seroprevalence in the Asia–Pacific region, only 25.1% [128]. People infected with HIV are at increased risk of suffering from severe forms of toxoplasmosis, such as cerebral toxoplasmosis. This condition has approximately a 30% mortality rate and may even leave surviving patients with lingering sequelae [9–12]. The Indonesian government recommends Toxoplasma screening for all patients newly diagnosed with HIV [129]. In addition, HIV-infected individuals with CD4 + T cells of fewer than 100 cells/µl and seropositive for Toxoplasma are required to receive primary prophylaxis via cotrimoxazole.

The connection between toxoplasmosis and psychiatric disorders in humans has been studied extensively. For example, Toxoplasma IgG seropositivity was associated with schizophrenia and bipolar disorder [130, 131]. Hence, psychiatric patients are among the populations of interest when studying toxoplasmosis. In our analysis, the estimated seroprevalence of toxoplasmosis among psychiatric patients was 63.21%. This result was higher than the estimated seroprevalence in Asia, which was 43.0% [132].

Regardless, our systematic review encountered several limitations. Most of the studies had a small sample size. Even when we set the standard to the modest 16.4% estimated prevalence in Asia overall [24], most studies did not meet our minimum sample size criteria. Studies with small-sample sizes tend to have wider margins of error, i.e., their observed prevalence is more likely to deviate from the true population prevalence [133]. In addition, most of the studies were conducted in Java. The number of studies conducted in other parts of Indonesia was scarce. Therefore, although we attempted to ensure the robustness of our findings, care should be taken when extrapolating the data to other regions. The final limitation that we encountered was our inability to access older studies. In total, we identified 20 records that could be included in our review. However, we could not find any digital or physical copies of these records.

Conclusion

Toxoplasmosis is highly prevalent among the human population in Indonesia, with an adjusted seroprevalence of 60.06%. However, our study mainly relied on studies with small sample sizes. Furthermore, most of the studies were performed in Java; therefore, some high-quality population-based studies must be conducted in other regions of Indonesia to better estimate the prevalence of toxoplasmosis across the country.

Abbreviations

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RoB

Risk-of-Bias

CI

Confidence Interval

LFK Index

Luis Furuya-Kanamori Index

PCR

Polymerase Chain Reaction

CSF

Cerebrospinal Fluid

ELFA

Enzyme-linked fluorescence assay

ELISA

Enzyme-linked immunosorbent assay

LAT

Latex Agglutination Test

LFA

Lateral-flow Immunoassay

ECLIA

Enhanced chemiluminescence immunoassay

IHA

Indirect Hemagglutination Test

MAT

Modified Agglutination Test

Authors’ contributions

TMP wrote the systematic review protocol, confirmed the eligibility of the included studies, cross-checked the validity of data extraction, analyzed the extracted data using R, and drafted the main manuscript text. AHD screened the eligibility of identified records, extracted the data of the included studies, and prepared the figures. SK screened the eligibility of identified records, extracted the data of the included studies, and prepared the additional files. AMTS assessed the quality of the included studies and prepared the tables. FPSW assessed the quality of the included studies and checked the reference list. All authors reviewed the manuscript.

Funding

This study received no funding.

Data availability

All data that were analyzed for this study are included in both the main manuscript and the additional files.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.