Rotavirus Prevalence and Genotypes Among Children Younger Than 5 Years With Acute Diarrhea at Mulago National Referral Hospital, Kampala, Uganda

Abstract

Background: 

Rotavirus is the most common cause of severe diarrhea among children <5 years of age worldwide. To determine the prevalence of severe rotavirus infection in children admitted with acute diarrhea attending Mulago National Referral Hospital in Uganda, active sentinel surveillance was conducted from July 2006 to December 2012.

Methods: 

A total of 6387 children with acute diarrhea were enrolled into the public health surveillance system and of these, 5627 had stool samples collected and tested for rotavirus antigens by enzyme immunoassay ProSpecT Rotavirus kit.

Results: 

Rotavirus was detected in 1844 (32.8%) of 5627 children with acute diarrhea that had stool specimens collected, and 93% of positive cases of rotavirus gastroenteritis were between 3 and 23 months of age, with highest prevalence in children 6–11 months of age. Rotavirus infections occurred throughout the year. During the surveillance period (2006–2012), a total of 354 positive stool samples were subjected to reverse transcription polymerase chain reaction and genotyping assays. The most common genotypes detected were G1P[8] (16.1%) and G9P[8] (15.3%), followed by G2P[4] (7.6%), G9P[6] (7.1%), G8P[4] (6.5%) and G12P[6](5.6%). Mixed G or P types (17.9%) and partially typed either G or P types (10.7%) were common.

Conclusion: 

Uganda would benefit by introducing rotavirus vaccine and hence reduce the hospitalization burden of managing acute diarrhea cases.

Worldwide, 10% of <5 mortality is caused by diarrheal disease.¹ Rotavirus is 1 of the 4 most common causes of moderate to severe diarrhea among children <5 years of age in Sub-Saharan Africa and Asia.² Diarrhea is 1 of the documented ‘10 killer conditions’ among Ugandan children with an estimated 10,637 deaths due to rotavirus occur annually in children < 5 years of age.³ Uganda Ministry of Health with support by the World Health Organization (WHO) established sentinel surveillance of severe rotavirus (RV) infections among children <5 years of age in Mulago National Referral Hospital (MNRH) in June 2006. Before the initiation of surveillance, local research on RV diarrhea in children was limited. Unpublished work (Kenya–Mugisha, 1989) from the 1980s found that rotavirus was responsible for 57% of acute watery diarrhea among the children studied. A cross-sectional study at MNRH in 2010 reported a prevalence of RV diarrhea of 45.4% (n = 340 cases).⁴ Among other factors, RV diarrhea was significantly associated with severe dehydration compared with non-rotavirus diarrhea cases. A case-control study in 2011 at MNRH found that breastfeeding was not protective of RV diarrhea.⁵ However, despite these studies and yearly reports of prevalence of RV diarrhea from this surveillance, there has not been a single published study that describes rotavirus disease trends spanning a period of >1 year in Uganda. Additionally, there has not been any documentation of the prevalent RV genotypes before 2006, when Uganda Ministry of Health and WHO established the first sentinel-based surveillance.

RV vaccine use in Sub-Saharan Africa has been limited to date. Sudan introduced RV vaccination in their national immunization program in 2011, the first Global Alliance for Vaccine Initiative-eligible country in Africa to do so,⁶ and several other Global Alliance for Vaccine Initiative-eligible countries in Africa are planning to introduce RV vaccine in 2012–2013. As of September 27, 2011, Uganda was 1 of the countries qualifying for (Global Alliance for Vaccine Initiative) RV vaccine support⁶ and is in a process of applying for this support to introduce the vaccine by 2015. A study that evaluated the potential impact and cost effectiveness of RV vaccination among the childern <5 years of age in Uganda found that introducing RV vaccine would save 5265 lives, prevent 94,729 cases of rotavirus diarrhea and save 996 million Uganda shilling annually.⁷

One of the main purposes of this surveillance is to determine the prevalence of RV diarrhea in MNRH and to establish the prevalent RV strains, before introduction of RV vaccination in Uganda.

MATERIALS AND METHODS

Study Site

MNRH is Uganda’s National Referral hospital and the main teaching hospital for Makerere University, College of Health Sciences. The Department of Pediatrics, 1 of the major departments in the hospital, treats approximately 25,000 children annually; of which, approximately 10,000 are managed as in-patients.

Case Definition

A suspected case of severe RV infection was defined in a child <5 years of age who was admitted for management of acute watery diarrhea and/or vomiting of <14 days duration.⁸ A confirmed case of RV diarrhea was any suspected case <5 years of age, whose stool RV was demonstrated by enzyme immunoassay (EIA). Therefore, the target population under surveillance was children admitted to the acute care unit who met the following inclusion criteria: <5 years of age, acute diarrhea, and /or vomiting, symptoms that lasted <14 days duration and admitted to acute care unit for emergency treatment of severe diarrhea. All cases with bloody diarrhea or children who acquired diarrhea while in the hospital were excluded from the study population.

Surveillance Methodology and Stool Collections

Sentinel-based rotavirus surveillance study at MNRH followed the WHO generic protocol and regional standard operating procedures (WHO-Regional Office for Africa).⁸ A line list of all suspected cases of rotavirus diarrhea was kept at the acute care unit ward.

Approximately 5 mL of stool specimens were collected from enrolled children within 24 hours of admission to the hospital. The specimens were immediately transported to the laboratory. Epidemiologic and laboratory data were collected from the logbook maintained in the children’s ward. The sentinel site regularly shared data on a monthly basis with Ministry of Health and WHO.

On an annual basis, a random sample of RV EIA-positive stool samples and 10% of RV EIA-negative samples were selected and sent to the WHO Rotavirus Regional Reference Laboratory in South Africa at the University of Limpopo (Medunsa Campus) for further tests and quality control. The Rotavirus Regional Reference Laboratory in South Africa performed genotyping of rotavirus-positive specimens and sent back results of genotyping to the national level and to WHO Regional Office for Africa.

LABORATORY METHODS

Rotavirus Detection

EIA ProSpecT Rotavirus kit (Oxoid Ltd, Hampshire, United Kingdom) was used to detect group A rotavirus. All details were followed as per manufacturer’s instructions. The results were read spectrophotometrically, the cut-off value was calculated by adding 0.200 to the negative control absorbance. Absorbance above the cut-off value was considered rotavirus positive.

Reverse Transcription Polymerase Chain Reaction and Genotyping Assays

At the Rotavirus Regional Reference Laboratory in South Africa, randomly selected rotavirus positives were further characterized. The rotavirus dsRNA was extracted from 10% stool dilutions using the QIAamp viral RNA extraction method (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions. The extracted viral RNA was reverse transcribed and amplified for VP4 and VP7 genomic segments using consensus primer sets Con2/Con3 and sBeg/end9.^9,10 The G and P genotypes were determined with seminested polymerase chain reaction amplification of VP7 and VP4 genes using a cocktail of human rotavirus primer sets as described.^9–13

RESULTS

Between July 2006 and December 2012, a total of 6387 cases of suspected severe rotavirus infection among children <5 years of age were hospitalized in MNRH. Of these 5627 (88.1%) had stool specimens collected and examined and 1844 (32.7%) were positive for rotavirus. The annual proportion of rotavirus positives ranged from 29% in 2011 to 40% in 2007 (Table 1). Rotavirus infection was common in 3–23 month olds with 1722 (93.3%) cases occurring in this age group. Boys accounted for 1127 (61.1%) cases. Vomiting was reported in 81.1% of rotavirus-positive cases and fever in 53.9%. Over 50% of the rotavirus cases had some or severe dehydration due to diarrhea and associated vomiting (Table 2). Approximately two-thirds of rotavirus-positive patients were treated with intravenous fluids. The median length of stay was 2 days. Two cases died.

TABLE 1: Enrollment of Children <5 Years of Age With Diarrhea and Results of Rotavirus Testing

TABLE 2: Characteristics, Clinical Symptoms, and Treatment of Children <5 Years of Age Hospitalized With Rotavirus and Non-rotavirus Diarrhea

Rotavirus infection occurs all year round with highest numbers seen during the rainy months of March through May and September through November (Fig. 1).

FIGURE 1: Rotavirus detection per month and year.

A total of 354 EIA rotavirus-positive stool samples were subjected to reverse transcription polymerase chain reaction and genotyping. Overall, the results indicated that the most predominant rotavirus strains detected were G1P[8] (16.1%) and G9P[8] (15.3%), followed by G2P[4] (7.6%), G9P[6] (7.1%), G8P[4] (6.5%), G12P[5.6%] and G1P[6] (4.2%). Many strains were detected as mixed G (11.3%) or mixed P (6.8%) types and partially G or P types (10.7%; Table 3).

TABLE 3: Summary of Rotavirus Genotypes Circulating in Uganda From 2007 to 2012

DISCUSSION

In Uganda, rotavirus infects young children and is a common cause of hospitalization due to severe disease. Our findings have shown that 95% of infected children are <2 years of age with the highest prevalence between 6 and 11 months old which is comparable with previous studies in Iran and Nigeria.^14–16 This onset of infection correlates very well with the decline of maternally acquired antibodies that disappear around 5 months. Many of the infections were severe with 83% of children experiencing at least some dehydration and 66% receiving intravenous rehydration. These children would benefit from a rotavirus vaccination program, a more cost effective intervention.

This surveillance found that rotavirus infections more commonly occurred in boys compared with girls. This observation has been previously documented.^4,17 However, other studies have shown no gender difference in rotavirus infection.¹⁸ It is possible that the apparent gender difference reflects a difference in health-seeking behavior with regards to children’s gender. Rotavirus is transmitted fecal orally. Rotavirus infections were detected year round in Uganda, which shows that rotavirus infection is a regular factor in the high disease burden among the study population. An estimated 10,637 children <5 years of age die in Uganda each year due to rotavirus diarrhea,³ underlining the importance of Rotavirus as a cause of child mortality in Uganda.

High rates of mortality have been documented in developing countries but in our study the mortality rate was low due to specialized care provided by the national referral hospital. This could be different in other rural hospitals in Uganda and many children may die from rotavirus diarrhea in the community without receiving medical care.

G1P[8] and G9P[8] were the most common genotypes detected in our study and are also the most common genotypes detected in studies globally (Table 3). However, the proportion of G1P[8] documented in our study (16.1%) is significantly smaller than that detected globally (52%).¹⁹ This difference could be due to the sampling strategy used to select specimens for genotyping in our study but also likely to reflect a true large diversity of circulating strains in Uganda. The 2 rotavirus vaccines in the market²⁰ (RotaTeq and Rotarix), which have demonstrated effectiveness against severe rotavirus diarrhea both in the developed and developing countries,²¹ cover the predominant genotypes that have been shown in our surveillance.

We had several limitations to this work. First, the surveillance was conducted at a single hospital, MNRH. As a national referral hospital, the patients enrolled in the surveillance may not be representative of patients with diarrhea at other hospitals in Uganda. Given the specialized care at MNRH, we almost certainly underestimated the mortality rate due to rotavirus diarrhea. Additionally, stool specimens were not collected from all children enrolled in the surveillance platform, which may have biased our estimate of rotavirus positivity. However, we obtained a stool specimen from 88% of enrolled children and the children without a stool specimen will likely have little impact on our overall findings. Finally, only a subset of rotavirus positives was genotyped, so the results may not be representative of all rotavirus-positive cases. However, the genotyping data do reflect the diversity of strains circulating in Uganda and Africa.

Public health surveillance is a useful tool that can be employed for decision making with regards to introduction of a new vaccine into the childhood immunization schedule as a response to high burden of disease. This study clearly shows that the proposed strategy for controlling rotavirus infection through routine vaccination will greatly reduce the workload, given the limited health workforce of managing children with severe acute diarrhea. More studies can be done from the public health surveillance system such as determination of risk factors so that a comprehensive approach is implemented to control rotavirus infection. Once rotavirus vaccine is introduced in Uganda, this surveillance system can serve as a platform for monitoring the impact and effectiveness of a national immunization program.

ACKNOWLEDGMENTS

The authors acknowledge WHO for material and technical support to the surveillance activity, which produced this work; all nurses and doctors at Acute Care Unit of MNRH; Laboratory staff both at MNRH and at the Department of virology, Medunsa Campus of Limpopo University, for their technical role in handling and analyzing study samples and all the parents of the children who consented to participate in the study.