Malaria prevalence and performance of diagnostic tests among patients hospitalized with acute undifferentiated fever in Zanzibar

Abstract

Background: Control efforts in Zanzibar reduced the burden of malaria substantially from 2000 to 2015, but re-emergence of falciparum malaria has been observed lately. This study evaluated the prevalence of malaria and performance of routine diagnostic tests among hospitalized fever patients in a 1.5 years period in 2015 and 2016.

Methods: From March 2015 to October 2016, paediatric and adult patients hospitalized with acute undifferentiated fever at Mnazi Mmoja Hospital, Zanzibar were included. The malaria prevalence, and performance of rapid diagnostic test (RDT) and microscopy, were assessed using polymerase chain reaction (PCR) as gold standard.

Results: The malaria prevalence was 9% (63/731). Children under 5 years old had lower malaria prevalence (5%, 14/260) than older children (15%, 20/131, p = 0.001) and persons aged 16 to 30 years (13%, 15/119, p = 0.02), but not different from persons over 30 years old (6%, 14/217, p = 0.7). All cases had Plasmodium falciparum infection, except for one case of Plasmodium ovale. Ten malaria patients had no history of visiting mainland Tanzania. The RDT had a sensitivity of 64% (36/56) and a specificity of 98% (561/575), and microscopy had a sensitivity of 50% (18/36) and a specificity of 99% (251/254), compared to PCR. The malaria parasitaemia was lower in patients with false negative results on RDT (median 7 × 10³ copies/µL, interquartile range [IQR] 2 × 10³ - 8 × 10⁴, p = 0.002) and microscopy (median 9 × 10³ copies/µL, IQR 8 × 10² - 7 × 10⁴, p = 0.006) compared to those with true positive RDT (median 2 × 10⁵ copies/µL, IQR 3 × 10⁴ - 5 × 10⁵) and microscopy (median 2 × 10⁵ copies/µL, IQR 6 × 10⁴ - 5 × 10⁵).

Conclusions: The study emphasizes that malaria was a frequent cause of febrile illness in hospitalized patients in Zanzibar in the years 2015-2016, particularly among school age children and young adults. We found evidence of autochthonous malaria transmission in Zanzibar. Compared to PCR, both RDT and microscopy had low sensitivity, and false negative results were associated with low parasitaemia. While low parasitaemia identified only by PCR in a semi-immune individual could be coincidental and without clinical relevance, clinicians should be aware of the risk of false negative results on routine tests.

Keywords: Eastern Africa; Fever; Malaria; Microscopy; Point-of-care diagnostic tests; Polymerase chain reaction; Prevalence; Surveillance; Tanzania; Zanzibar.

Fig. 1

Patients included and analyses performed

Fig. 2

An overview and results from the analyses performed by PCR, RDT and routine microscopy. The numbers in the circles refer to malaria positive results performed by each method. The numbers of malaria negatives are given in the bottom of the squares. Except for the malaria prevalence, all numbers are given for the performance of each method independently of false positives/negatives by the gold standard method PCR

Fig. 3

Malaria parasitaemia by age group (years) and result of microscopy and malaria rapid diagnostic test (RDT). Quantitation of parasitaemia by real-time PCR by diagnostic modalities expressed as log-transformed values of copies per µL blood. Unit of measurement for parasitaemia by PCR is described in the section PCR methods. Dots represent individual observations. Number tested in brackets. * Kruskal Wallis test † Wilcoxon rank sum test

Fig. 4

Malaria cases by age groups. Number of malaria patients (blue bars) among febrile patients (grey bars) and percentage positive (line) in different age groups

Fig. 5

Monthly number of malaria cases and monthly rainfall from February 2015 to October 2016. Rainfall data for Dar es Salaam from the Tanzanian Meteorological Agency (TMA) [35, 36]. Study start March 17, 2015, study end October 4, 2016