Seroprevalence of SARS-CoV-2 in Tanzania: results from a cross-sectional study documenting the Omicron wave in five regions of Tanzania

Abstract

Background

Between August and October 2021, before the Omicron wave and before COVID-19 vaccines became widely available in Tanzania, two studies estimated that half of the population in Zanzibar and Mwanza in Tanzania had SARS-CoV-2 antibodies. From May 2020 to Dec 2021, Tanzania adopted a very different COVID-19 mitigation strategy with no lockdowns or quarantines and limited community-level COVID-19 testing. We conducted population-based SARS-CoV-2 seroprevalence surveys in five regions of Tanzania between November 2021 and July 2022 to estimate community-level exposure to SARS-CoV-2and vaccine coverage.

Methods

Regionally representative cross-sectional serosurveys were progressively conducted via multistage cluster sampling in five regions in Tanzania. Twenty enumeration areas were randomly selected from each region. Seropositivity was determined using the WANTAI SARS-CoV-2 Ab ELISA kit.

Results

We found that nine out of every ten people living in five geographically dispersed regions in Tanzania had SARS-CoV-2 antibodies despite only 8% vaccination coverage in our sample.

Conclusion

Given the high seroprevalence of SARS-CoV-2 in these regions, targeting COVID-19 vaccination efforts to those at greater risk of severe disease may help maximize the public health impact.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-26079-5.

Introduction

Reported COVID-19 cases to the World Health Organization (WHO) grossly underestimated true infections and likely reflected only symptomatic cases. Early in the COVID-19 pandemic (March 2020–August 2020), Tanzania swiftly implemented a range of protective measures, but stringent mitigation subsequently waned, and the reporting of cases to the WHO was suspended for a prolonged period [1].

Globally, as of August 2023, over 769 million confirmed cases and over 6.9 million deaths were reported to the WHO [2]. As of May 2023, Tanzania reported 43,078 cumulative total confirmed cases, and 841 deaths (2% of all reported cases) to the WHO [3]. These numbers of cases and deaths may be underestimated and likely reflect symptomatic or laboratory-confirmed COVID-19 cases [4], as surveillance efforts in Tanzania have focused primarily on data from travelers or patients with severe illness who seek medical care. Therefore, the extent of the epidemic in Tanzania, considering the presence of mild and asymptomatic infections that do not necessitate medical attention, has remained largely unknown [5].

Many countries’ responses to the COVID-19 pandemic included serologic surveillance studies in addition to increasing molecular testing for the detection of active disease. SARS-CoV-2 antibodies are detected via serological testing, which in unvaccinated individuals indicates prior infection with SARS-CoV-2 [6]. Population-based serological testing complements diagnostic testing for acute infections by providing estimates of the seroprevalence and can be used to inform public health response strategies to control the COVID-19 pandemic [7, 8]. Moreover, synthesizing sero-epidemiological findings is crucial in tracking the transmission of infection, identifying disproportionately affected groups, and monitoring the development of herd immunity in this era of vaccines and new highly infectious variants, such as Omicron [9, 10].

During the first year of the COVID-19 pandemic, most SARS-CoV-2 seroprevalence studies were carried out in high-income countries, leaving a substantial evidence gap in low- and middle-income countries (LMICs). In Tanzania, two prior seroprevalence studies were conducted in the Mwanza region and Zanzibar region from August–September 2021 and reported 50–57% SARS-CoV-2 positivity [11, 12]. Findings from these two studies provided some insights on SARS-CoV-2 transmission in Tanzania, however, both studies had limited geographical focus and were conducted before the emergence of the Omicron variant, therefore unsuitable for providing a snapshot of the regional diversity in Tanzania. A study among mother‒baby pairs in the Tanga region reported SARS-CoV-2 seroprevalence estimates of 29%among mothers and 40% among infants [13]. Sero-epidemiological dynamics of SARS-CoV-2 in Tanzania are especially important considering the differences in response dynamics compared with those in neighboring East African countries, where severe mitigation measures, including prolonged lockdowns and quarantine, have been enforced [2]. This study aimed to determine the seroprevalence and sociodemographic differences of SARS-CoV-2 infection in five regions from five distinct geographical zones in the Tanzania mainland to inform public health actions and policies related to the COVID-19 response.

Methods

Study design and settings

Regionally representative cross-sectional serosurveys were progressively conducted via multistage cluster sampling in five regions in Tanzania. The regions were purposively selected from mainland Tanzania with considerations to include (i) regions with low and high burdens of COVID-19 hospitalization, (ii) at least one region from each zone of the country, and (iii) logistical reasons such as the feasibility of sample handling and transportation. With these considerations, we purposively selected Dar es Salaam, Arusha, and Mwanza (high burden of COVID-19 hospitalizations) and Mtwara and Iringa (low burden of COVID-19 hospitalizations) from the southern zone of the country. In Mtwara region, two cross-sectional surveys (one in November 2021 and another in February 2022) were conducted to capture the serological profiles before and after the emergence of the Omicron variant which was first reported in South Africa on November 24, 2021.

Sampling

Within each region, 20 enumeration areas (EAs) were randomly selected and stratified into urban and rural/semiurban areas. The 2012 Population and Housing Census data provided the sampling frame for selecting EAs using probability proportions to size sampling. EA sampling was performed separately in urban and rural settings. The allocation of the number of EAs selected in the respective settings was based on the level of urbanization of a particular region; i.e., the higher the level of urbanization was, the greater the share of EAs selected in the urban setting. Within each selected EA, 20 households were randomly sampled, and participants were approached for informed consent. Participants aged 18 years and above were enrolled after providing written consent to participate. For those aged below 18 years, written consent was sought from parents/legal guardians in addition to assent from children aged 13–17 years. Whole blood samples were collected from participants aged 5 years and above only. Children aged less than 12 months were excluded. Individuals who declined to provide written consent, or lived in communal/institutionalized residences such as boarding schools, prisons and refugee camps were also excluded.

The Mtwara region was the first region to be surveyed in November 2021. By mid-December 2021, with the highly contagious new Omicron variant sweeping across East Africa [5], we modified the sampling strategy to repeat the survey on Mtwara (in February–March 2022) to understand the potential effects of Omicron on SARS-CoV-2 seropositivity.

Sample size Estimation

We explored various scenarios for estimating the required sample size that considered the assumed prevalence of SARS-CoV-2 antibodies in the population, the number of participants selected from a cluster, the intraclass correlation coefficient (ICC), the desired precision (margin of error), and the α level. In the final sample size calculation for each region, we assumed a SARS-CoV-2 antibody prevalence of 20% on the basis of a published pooled analysis of prevalence reported in sub-Saharan Africa [1]; a margin of error of +/−4.9%; a design effect of 3 (based on ICCs of 0.05 and 55 participants from each cluster); and an α value of 0.05. We assumed a conservative response rate of 70%, given the uncertainties around community willingness to participate in such a seroprevalence survey. On the basis of these assumptions and parameters, we estimated that a minimum sample of 1095 individuals from each region would be needed for a margin of error of < 5%.

Recruitment and specimen collection

Android tablets programmed with the Open Data Kit (ODK) application were used to collect individual questionnaire (Appendix 1) information on sociodemographic characteristics and self-reported COVID-19 vaccination status. After the interview, 5 mL of venous blood was collected from each participant aged five years and above into BD vacutainer blood collection tubes. The blood samples were processed on the same day at a nearby health facility/laboratory by centrifuging at 3000 rpm for 10 min to obtain serum, which was stored in cryovial tubes at − 20 °C (− 4 °F) and then packed and labelled in a proper triple package and transported in a cold chain to the National Public Health Laboratory (NPHL) in Dar es Salaam for testing. Seropositivity was determined via the WANTAI SARS-CoV-2 Ab ELISA (Beijing Wantai Biological Pharmacy Enterprise Co. Ltd) following the manufacturer’s recommendations [14]. Manufacturer validation reports indicate the kit sensitivity of 55.38% (≤ 7 days), 84.78% (8–14 days), and 98.7% (≥ 15 days), and specificity of 98.6%. The combined PPV was 67.1% (33.6% −88.4%), and the combined NPV was 99.8% (99.0% − 100.0%) [15]. ELISAs detect IgG levels in individuals infected with SARS-CoV-2 or who have received COVID-19 vaccination. The U.S. Centers for Disease Control and Prevention reviewed the assessment and determined it to be a non-research public health activity. Tanzania’s National Health Research Ethics Committee also reviewed and approved the evaluation (Ref. No NIMR/HQ/R.8a/Vol.IX/3706).

Analytic methods

In all the visited regions, enumeration areas and household and individual population weights were created to account for sample selection probabilities and nonresponses. For region-specific estimates, the region weights were used in the weight adjustments, and for the pooled analyses (for combined region estimates), the pooled weights accounting for region weights were used in the weighted adjustments. Seropositivity for SARS-CoV-2 was defined as positive samples on the ELISA. Descriptive analyses, i.e., proportions with their respective 95% intervals, were calculated while accounting for the complex survey design and weighted adjustments. The proportions disaggregated by age, sex, region, location (rural vs. urban), COVID-19 vaccination status (self-reported), and date of specimen collection were also reported. The p values from design-based F statistics were used to test the associations between the categories for data that did not exhibit trends. For the data that exhibited a trend, i.e., comparing proportions across the months of data collection, the Cochran‒Armitage test for trend was used to test the associations. P values of < 0.05 were considered statistically significant. The analyses were conducted via Stata 18 (College Station, TX).

Results

A total of 2,220 households were randomly selected to participate in the study. Among the selected households, 1,995 (90%) were reached after 3 visits. Among the 1,995 households, 1,601 (80%) consented to participate in the study. The data collection teams approached 4,862 participants from 1,601 consented households across the five regions between November 15, 2021, and July 24, 2022 (Fig. 1). A total of 4,480 (92.1%) participants provided consent and were enrolled in the study, and 3,815 (85.1%) participants provided blood samples, which were tested via ELISA (Fig. 2). The analytical sample comprised 3,815 participants whose available seropositivity test results were determined via an enzyme-linked immunosorbent assay (ELISA). Overall, the seroprevalence of SARS-CoV-2 was 88% (95% confidence interval [CI]: 86%−90%). The seroprevalence was lower in Mtwara before Omicron (75%; 95% CI: 70–79) than after Omicron (92%; 95% CI: 89%−95%). We did not detect any significant differences in seroprevalence by age or sex. Among the 3,815 participants with ELISA results, 311 (8%) self-reported that they had been vaccinated against COVID-19. The percentage of seropositive participants was 89% (95% CI: 86%−90%) among unvaccinated participants and 94% (95% CI: 89%−97%) among vaccinated participants (p = 0.049). The prevalence of SARS-CoV-2 antibodies increased from the lowest recorded value of 75% (95% CI: 70%−79%) in November 2021 to 99% (92%−100%) in July 2022 (Table 1). Region-specific estimates of the seroprevalence of SARS-CoV-2 are shown in Table 2.

Fig. 1.

Timeline for data and sample collection for SARS-CoV-2 seroprevalence surveys, Tanzania, 2021-2022

Fig. 2.

Flow chart describing the number of households and participants approached, enrolled, or excluded, SARS-CoV-2 seroprevalence survey, Tanzania 2021–2022

Table 1.

Characteristics and serologic results of individuals whose serum samples were tested for the presence of SARS-CoV-2 antibodies, Tanzania, 2021–2022

Characteristic	Total	Positive (N)	% (95% CI) (Weighted)	P value
Overall	3815	3,441	88 (86–90)	-
Age group (Years)
5–17	693	606	87 (82–90)	0.232γ
18+	3,097	2816	89 (87–91)
Sex
Female	2154	1961	89 (86–91)	0.539γ
Male	1661	1480	88 (84–90)
Region
Mtwara (Phase 1) (Nov 15–25, 2021)	483	353	75 (70–79)	< 0.001γ
Mtwara (Phase 2) (25 Feb 2022-27 March 2022)	680	633	92 (89–95)
Iringa (05 April 2022-13 May 2022)	874	815	92 (90–94)
Arusha (February 3, 2022-July 24, 2022)	633	613	98 (96–99)
Dar es salaam (Dec 01–18, 2021)	389	334	84 (78–88)
Mwanza (05 April 2022-09 May 2022)	756	693	92 (90–94)
Location
Rural	2356	2091	89 (87–90)	0.612γ
Urban	1459	1350	88 (84–91)
Vaccination status (adults 18 + years)
Vaccinated	311	300	94 (89–97)	0.049γ
Unvaccinated	2798	2524	89 (86–90)
Dates of specimen collection
November 2021	482	352	75 (70–79)	< 0.001φ
December 2021	389	334	84 (78–88)
February 2022	135	129	96 (91–99)
March 2022	1099	1039	95 (93–96)
April 2022	1222	1140	92 (90–94)
May 2022	409	369	92 (88–95)
July 2022	79	78	99 (92–100)

γ = P values from the design-based F statistic, φ = P values from the Cochran‒Armitage test for trend

Table 2.

Region-specific estimates of the Seroprevalence of SARS-CoV-2 antibodies in Tanzania from 2021–2022

Variable	Total n	Positive n	% (95% CI) (Weighted)	P value γ
Dar es Salaam Region	389	334	84 (78–88)
Sex
Female	225	196	85 (77–90)	0.635
Male	164	138	83 (72–89)
Age group
5–17	88	73	82 (69–90)	0.682
18+	299	259	84 (77–90)
Vaccination status
Vaccinated	25	22	87 (61–97)	0.755
Unvaccinated	274	237	84 (77–90)
Arusha Region	633	613	98 (96–99)
Sex
Female	371	358	97 (95–99)	0.795
Male	262	255	98 (95–99)
Age group
5–17	78	75	98 (93–99)	0.767
18+	552	535	97 (95–98)
Vaccination status
Vaccinated	76	76	100	0.228
Unvaccinated	476	459	97 (95–98)
Mtwara region (1st visit)	483	353	75 (70–79)
Sex
Female	275	205	76 (70–82)	0.465
Male	208	148	73 (65–80)
Age group
5–17	102	74	73 (62–82)	0.636
18+	367	269	75 (70–80)
Vaccination status
Vaccinated	42	37	88 (70–96)	0.077
Unvaccinated	337	240	74 (68–79)
Mtwara Region (2nd visit)	680	633	92 (89–95)
Sex
Female	358	341	94 (90–97)	0.131
Male	322	292	90 (83–94)
Age group
5–17	122	110	87 (76–94)	0.016
18+	557	523	95 (93–97)
Vaccination status
Vaccinated	81	81	100	0.035
Unvaccinated	476	442	94 (91–96)
Mwanza Region	756	693	92 (90–94)
Sex
Female	436	399	90 (86–94)	0.110
Male	320	294	94 (91–96)
Age group
5–17	196	177	90 (84–94)	0.076
18+	559	515	94 (91–96)
Vaccination status
Vaccinated	39	37	94 (80–99)	0.724
Unvaccinated	520	478	95 (91–96)
Iringa Region	874	815	92 (90–94)
Sex
Female	489	462	94 (91–96)	0.133
Male	385	353	91 (86–94)
Age group
5–17	107	97	90 (82–95)	0.218
18+	763	715	94 (91–95)
Vaccination status
Vaccinated	48	47	98 (86–100)	0.207
Unvaccinated	715	668	93 (91–95)

γ=Pvalues from the design-based F statistic

Discussion

Our study revealed high SARS-CoV-2 seropositivity (88%) across five geographically diverse regions in Tanzania, with seropositivity increasing over time from 75% in November 2021 to 99% in July 2022. With the exception of November 2021, most of the studies’ data were collected around or after the emergence of the Omicron variant in the subregion and before the COVID-19 vaccination among the general population was scaled up. Despite the possibility of waning antibodies [16], our study revealed that widespread SARS-CoV-2 infection had been experienced in the Tanzanian population even before the Omicron variant but was likely further hastened by the emergence of this variant. The high seroprevalence detected was almost entirely attributed to previous SARS-CoV-2 infections, as evidenced by the 89% prevalence of SARS-CoV-2 antibodies among participants who had not received a COVID-19 vaccine. This can be attributed to the relaxed mitigation measures taken by the country during the pandemic [1]. However, the ELISA kits we used could not differentiate between infection-induced and vaccination-induced antibodies. The high prevalence findings from Tanzania are comparable to those observed in Kenya at a similar time point during the pandemic [17]. Another seroprevalence study conducted in Ghana in December 2021 reported a high seroprevalence of 86.8% [95% CI: 84.53–88.77], attributed to the omicron wave. This suggests widespread exposure to the omicron variant across African population [18].

To our knowledge, only three prior SARS-CoV-2 seroprevalence studies have been conducted in Tanzania, all before the emergence of the Omicron variant, each with a limited geographic focus. Studies conducted in Zanzibar between August and October 2021 reported a seroprevalence of 57% [11]; the Mwanza region in September 2021 reported 50% [12], and the Tanga region reported 40% seroprevalence [13]. By the time of our first round of surveys in Mtwara in November 2021, a much higher seroprevalence (75%) was recorded than that previously reported in the Mwanza and Zanzibar studies [11]. However, earlier/historical seroprevalence data for Mtwara are not available for comparison with the periods of data collection from Mwanza and Zanzibar studies. Although plausible, it is unclear whether the Omicron variant alone accounted for the higher seroprevalence detected in our study than in previous ones [5]. However, findings from another study that conducted SARS-CoV-2 genomic surveillance in the country confirms the predominance of omicron variant in these regions around the same time [19].

The findings of this study are subject to limitations. First, the fast-changing epidemiological dynamics cannot be captured with the current cross-sectional design. Second, we only had pre-Omicron data for one of the five regions sampled. Finally, although five demographically diverse regions are included in the study, the findings are not nationally representative. Nevertheless, they represent the major geographical zones of the country. Other regions in Tanzania may have different seroprevalence rates than those reported in these five regions.

This study provides the first estimates of SARS-CoV-2 seroprevalence in multiple regions in Tanzania, while considering the COVID-19 vaccination coverage. Data from Mtwara region provide the first estimates of SARS-CoV-2 seroprevalence in Tanzania before and after Omicron wave.

Conclusion

The findings from the present study illustrate that a large proportion of participant in these five regions had been exposed to SARS-CoV-2 at some point, with no significant differences according to age or sex. The study suggests that current data on COVID-19 reported from Tanzania are representative of the situation in 2020 during the pandemic when Tanzania stopped reporting cases regularly [3]. Alternative methods such as reviews of excess mortality, changes in acute respiratory illness disease burdens or more focused modelling may be needed to gauge disease spread in Tanzania. Nevertheless, targeting vaccination among subgroups at greater risk of severe COVID-19 may maximize the public health benefits of COVID-19 vaccination in Tanzania.

Abbreviations

COVID-19

Coronavirus Disease 2019

EA

Enumeration Area

ELISA

Enzyme-Linked Immunosorbent Assay

ICC

Intraclass Correlation Coefficient

LMIC

Low- and Middle-Income Countries

NPHL

National Public Health Laboratory

ODK

Open Data Kit

SARS-CoV-2

Severe Acute Respiratory Syndrome Coronavirus 2

WHO

World Health Organization

Authors’ contributions

SM, GK, AW, NM, LS, CL, JK, SM, EN, MFJ, GJ, PA, MK, GS, MS, TN, NK, and WG designed the study and directed its implementation, including quality assurance and control. SM, GK, AW, NM, LS, CL, JK, SM, EN, MFJ, GJ, PA, MK, GS, MS, TN, NK, and WG designed the study and directed its implementation, including quality assurance and control. SM, GK, SM, EN, and AS supervised the field activities. SM, AW, LS, SM, MFJ, PA, SM, JD, CN, NK, and WG designed and carried out the study’s analytic strategy. LJ, MB; performed laboratory testing. All the authors contributed to the drafting, review, commenting, and approval of the manuscript. All the authors have read, approved, and declared that they have no conflicts of interest.

Funding

We acknowledge the following institutions that provided funding support: The U.S. Centers for Disease Control and Prevention (5 NU2HGH000054-02-00); and the US National Institutes of Health (Grant # U01AI151799), through the Center for Research in Emerging Infectious Diseases – East and Central Africa (CREID-ECA).

The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data availability

The data that support the findings of this study are available from the National Institute for Medical Research, Dar es Salaam, Tanzania, but the data transfer agreement applies to the availability of these data, which were used under licence for the current study and are not publicly available. Data are, however, available from the authors upon reasonable request and with permission of the National Institute for Medical Research, Dar es Salaam, Tanzania.

Declarations

Ethics approval and consent to participate

The research was conducted in compliance with the Declaration of Helsinki, and ethical approval was obtained from the Tanzania National Health Research Ethics Committee (NatHREC with Ref. No NIMR/HQ/R.8a/Vol.IX/3706). Written informed consent was obtained from all participants 18 years of age and older, and assent (7–17 years) and parental or guardian written informed consent and assent were obtained from participants younger than 18 years.

Consent for publication

The consent for publication was included in the informed consent process, and permission to publish was obtained from the NaTHREC.

Competing interests

The authors declare no competing interests.