Horizontally-acquired HIV infection in Kenyan and Swazi children

Abstract

Despite many reports of HIV-infected African children who have HIV-uninfected mothers, little is known about the extent and modes of horizontal HIV transmission in African children. We estimated the extent of horizontal HIV transmission in Swazi children by comparing child and mother HIV statuses in the 2006–2007 Swaziland Demographic and Health Survey (DHS). To identify correlates of horizontal HIV transmission, we conducted a case-control study of Kenyan children with horizontally acquired HIV infections and their uninfected siblings. Of 50 HIV-positive Swazi children in the DHS, 11 (weighted percent = 20, 95% confidence interval 11–33%) had HIV-negative mothers. These 11 children represented 0.6% of all Swazi children aged 2–12 who lived with their mothers. In the Kenyan study, children with horizontally acquired HIV infections had more kinds of blood exposures than their uninfected siblings. In particular, punctures related to health care for suspected malaria (phlebotomy, injection and infusion), injections while hospitalized and dental surgery (especially by informal providers) were more common in infected children. Horizontal HIV transmission appears to be common in some sub-Saharan African countries, and blood exposures seem to be the most likely routes of transmission. Rigorous surveillance and investigation of horizontally acquired HIV infection in children are urgently needed, along with universal public education about risks of specific blood exposures and ways to avoid them.

Keywords

HIV Africa iatrogenic disease matched sibling case-control study

INTRODUCTION

Since 1984, clinicians and researchers have repeatedly reported cases of horizontally-acquired HIV infections in African children, as indicated by infected children who have HIV-negative mothers.^1–12 On a population level, horizontal HIV transmission in children may be significant in some countries. For instance, in a 2005 national serosurvey, Rehle et al. ¹³ indirectly estimated the annual incidence of HIV infection to be 0.5% in South African children aged 2–14.

By investigating correlates of HIV infection in children who have HIV-negative mothers, epidemiologists have sought to identify possible routes of horizontal transmission. In Kinshasa, Democratic Republic of Congo (then Zaire) during 1984–1985, medical injections, blood transfusion and hospitalization were associated with HIV infection in paediatric patients aged 2–24 months.¹⁴ In 2004, Shisana et al. ¹ assessed several kinds of blood and breastmilk exposures in children recruited from primary health clinics in the Free State and also tested the children and their biological mothers for HIV infection. In children of HIV-negative mothers, injections at a dental care facility, surrogate breastfeeding and feeding with expressed milk from a hospital milk room were associated with HIV infection.

We report two new studies of horizontal HIV transmission in children living in countries (Swaziland and Kenya) not covered in prior research on the topic. The Swazi study is based on a national representative sample and allows estimation of the prevalence of horizontally-acquired HIV infection in children. In the Kenyan study, we employed a matched case-sibling control design and assessed potential blood exposures to HIV more comprehensively than heretofore.

METHODS

Swaziland 2006–2007 demographic and health survey

We analysed publicly available data from the 2006–2007 Swaziland Demographic and Health Survey (DHS; www.measuredhs.com).^15,16 The Swaziland DHS was a large, standardized, national probability sample household survey of persons aged 15–49 years. In half of the sampled households, all persons aged 12 and older were eligible for interview, and all persons aged two and older were eligible for anonymous HIV testing conducted by DHS staff. Overall, 95% of sampled households participated, and in these households, 94% of the eligible women aged 15–49 responded. HIV testing of dried blood spots was performed for 87% of the women aged 15–49 and 88–91% of children aged 2–14 in participating households.

In the DHS interview, mothers reported various details about each child they had borne. We infer that all such children were the biological offspring of the mothers. We included in our analyses all 1655 mother–child pairs in which both mother and child lived in the same household and were tested for HIV by the DHS staff and in which the child was between two and 12 years of age. (Approximately one-third of Swazi children in the DHS did not live with their mothers.¹⁵) We assumed that children in this age range had never engaged in consensual sexual intercourse. HIV serostatus was determined with an enzyme-linked immunosorbent assay (ELISA) test and confirmed by a different ELISA test; specimens with repeatedly discrepant ELISA results were tested by Western blot for final determination of serostatus.

We computed the unweighted and weighted proportion of HIV-positive children aged 2–12 who had HIV-negative mothers, and thus were presumably infected through horizontal modes of transmission. The weighted estimate accounts for differential probabilities of being sampled (based on the sampling design) and variations in response rates across demographic groups. We also computed the unweighted and weighted proportions of Swazi children overall who had horizontally-acquired HIV infection. To obtain a rough estimate of the number of Swazi children with horizontally-acquired HIV infection, we multiplied the weighted prevalence estimate by the number of Swazi children between age 0 and 14 in 2005,¹⁷ and reduced the product by 1/3 (the approximate proportion of Swazi children in the DHS not living with their mothers).¹⁵ We calculated univariate summary statistics to describe the HIV-infected children who had HIV-uninfected mothers in terms of age, sex, family wealth and urban/rural residence. We also explored associations between potential blood exposures to HIV from vaccination (bacille Calmette-Guérin [BCG], diphtheria, pertussis and tetanus [DPT], hepatitis B vaccine [HPV]) and measles) and HIV status in children who had HIV-negative mothers. Vaccination data were reported only for children under age five at the time of the survey.

Case-control study in Kenya

Between March 2007 and March 2008, we contacted 19 clinical staff (medical doctors, nurses, laboratory technicians and social workers) who worked in hospitals and clinics across Kenya, particularly in Nyanza Province and Nairobi. We asked these health-care providers to refer to us cases of HIV in children who had HIV-negative mothers. Parents of all participating children voluntarily consented to referral and participation in the investigation of their children's infections. Although providers reported to us that they had encountered more paediatric cases of horizontally-acquired HIV infection than they had referred (some whose families refused to participate and others who were not invited to participate by the providers), we have no data on the number of such unreferred cases. We also have no data on the consistency with which mothers of paediatric HIV cases diagnosed in these facilities were tested for HIV. Providers were not compensated in any way for referrals.

In April 2008, we interviewed the HIV seronegative mothers of 11 infected children under age 11. Interviews were conducted in Luo, Swahili, English, Luhya or Kisii. Each family was paid 500 Kenyan shillings (approximately US$8) after the interview to compensate for their time. Parents were not told before or during the interviews that they would receive such payment, so these payments did not act as incentives to participate. In our analysis, we included eight of these infected children who met the following criteria: documentation of mother's biological maternity of the infected child, documentation of mother's and infected child's serostatuses, no reported sexual exposures of any kind for the infected child, no reported surrogate breastfeeding of the infected child and an uninfected child sibling who met the first two criteria. Acceptable forms of documentation of biological maternity included the mother's national identification card and the child's birth certificate, vaccination record, baptismal card or the local chief's birth records (each of these documents for the child include the birth mother's name). Documentation of mother and child HIV serostatuses consisted of printed test results from known clinics, voluntary counselling and testing centres, or hospitals. In all cases, rapid HIV tests were used, and positive results were confirmed by a second rapid HIV test of the same blood specimen. We selected the sibling closest in age to the infected child (case) as a control. Of the three cases excluded from analysis, one had been breastfed by a woman other than her mother and whose serostatus was unknown, and two had no siblings with documented HIV test results.

In the interviews, we assessed the child's and parents' demographics and HIV testing histories, as well as the child's birth complications, breastfeeding experiences, blood and sexual exposures, hospitalization, malaria treatment and medication history. We also asked mothers and fathers how they believed their children acquired HIV. For most potential blood exposures we assessed the corresponding reason, location, date, provider and perceived hygiene for particular exposure episodes. Some parents did not report on every episode of a given exposure in such detail, and therefore we report on episodes only in a limited way and do not focus on exposure frequency. Any reported immunization (BCG, measles, DPT, haemophilus influenzae, HPV and/or yellow fever) was assumed to have occurred in a child's first year of life when the date of immunization was not specifically reported or documented. Unfortunately, we assessed receipt of injections only for those administered for malaria treatment, while hospitalized, for anaesthesia during dental or medical surgery, or by informal providers. We did not assess some potential blood exposures effectively, including infusions for reasons other than malaria treatment, wound care and physical contact with another's blood from a bloody nose, child birth or wound.

In analysis, we classified cases as exposed only if a particular exposure occurred before HIV diagnosis. If the date of earliest exposure or date of first positive test was reported as a year only and the other date (of earliest exposure or first positive test) was reported as occurring in the same year, then the case was classified as unexposed. We coded exposure in a parallel fashion for controls. We computed the diversity of blood exposures (number of different types) experienced by each child as a summary measure of blood exposures. Because the symptoms of malaria and primary HIV infection are sometimes similar and because HIV infection might increase susceptibility to malaria,^18,19 we also computed a diversity measure of blood exposures excluding those related to malaria treatment.

For each blood exposure, we computed the percentages of cases and controls exposed. We measured the magnitude of differences between cases and controls on the measures of blood exposure diversity with the point biserial correlation coefficient and assessed the significance of the differences with matched-pair t-tests and Wilcoxon-signed rank tests.

RESULTS

Swaziland 2006–2007 DHS

One thousand six hundred and sixty-five children aged 2–12 were tested for HIV and lived with their mothers who were also tested for HIV. Fifty (3.0%) of these children were HIV positive. Of the 50 seropositive children, 11 (unweighted: 22%, 95% confidence interval [CI] 13–35%; weighted: 20%, 95% CI 11–33%) had seronegative mothers. These 11 children represented 0.6% (weighted; 0.7% unweighted) of all Swazi children aged 2–12 who lived with their mothers. This prevalence of horizontally-acquired HIV infection in children is roughly equivalent to 1808 (0.006 × 449,869 × 0.67) Swazi children who have horizontally acquired HIV infection and lived with their mothers. Six of the 11 children with horizontally-acquired HIV infection in the DHS were girls and five were boys. They ranged in age from two to 12 years (median = 5) and ranged from the second lowest to the highest quintiles of family wealth (median = middle quintile). Two of these children lived in urban areas and the other nine lived in rural areas. Only three of the 11 children were under age five, and they had all received each of their BCG, DPT, HBV and measles vaccinations.

Case-control study in Kenya

Many mothers, cases, and controls had repeated HIV testing episodes to confirm their serostatuses. All mothers had tested HIV seronegative at least once since their child had been diagnosed with HIV (three mothers had two documented HIV seronegative test results). All controls had tested HIV seronegative at least once since their sibling had been diagnosed with HIV (3 controls had two seronegative test results, and two controls had three seronegative test results). All cases had documented positive test results; two cases had two positive test results, three cases had three seropositive test results and one case had four seropositive test results. Also, fathers of four child pairs had tested for HIV (all seronegative) since their infected children had been diagnosed with HIV.

Most cases were male and most controls were female, and most participating families resided in rural areas (Table 1). Cases were diagnosed between age four and 10. Exposure periods (time between birth and diagnosis [for cases]/last negative test [for controls]) were 23% longer, on average, for controls, although cases and controls were similar overall in age. Most children's parents were married and had secondary or higher education, and most cases resided in Nyanza Province. The reasons why cases initially tested for HIV (i.e. when diagnosed) were: medical advice in response to medical history and/or current symptoms (n = 5), universal screening at school (n = 1), presurgical assessment (n = 1) and school medical exam (n = 1).

Table 1

Comparison between Kenyan children with horizontally acquired HIV infection and their uninfected siblings, 2008

	Sibling pairs
Variable	Cases (n = 8)	Controls (n = 8)
Demographics
Age at interview, mean (SD)	9.7 (8.1–11)	9.7 (6.2–14)
Age at diagnosis (cases)/last test (controls), mean (range)	7.4 (4.7–10)	9.1 (6.1–14)
Sex (%) male	75	38
Rural residence (%)	75	75
Province
Nyanza (%)	63	63
Central (%)	25	25
Eastern (%)	13	13
Parents married (%)	63	63
Highest level education of mother or father, secondary or more (%)	75	75
Exposures (%)
Dental surgery (formal or informal providers)	38	13
Incisions by traditional healer (including bloodletting/saro)	38	38
Infusion during malaria treatment*	38	13
Injection during malaria treatment*	88	13
Injection while hospitalized^†	50	0
Injection from informal provider	25	25
Jigger (chigoe or sand flea) removal	38	13
Phlebotomy (for reasons other than HIV testing)	63	25
Piercing (ear)	13	38
Surgery (by formal or informal provider)	13	0
Vaccination injection	75	75
Other sharp (accidental needlestick)	13	0
Diversity (number of types) of exposures ^‡
Mean (SD)	5.0 (1.1)	2.5 (0.9)
Median (range)	5 (3–6)	2.5 (1–4)
Diversity of exposures excluding invasive malaria treatment ^§
Mean (SD)	3.4 (0.7)	2 (1.1)
Median (range)	3.5 (2–4)	2 (1–4)

SD = standard deviation

Note: Percentages do not always sum to 100 due to rounding

*From formal health-care provider

^†All hospitalization episodes occurred after the child had been weaned

^‡Difference between cases and controls: point biserial correlation coefficient = .80, matched pair t-test P < 0.01, Wilcoxon-signed rank test P < 0.05

^§Excludes any potential blood exposure related to diagnosis or treatment of malaria; difference between cases and controls: point biserial correlation coefficient = .59; matched pair t-test and Wilcoxon-signed rank test, P < 0.05

Cases had substantially and significantly more types of blood exposures than controls on average (Table 1). Of the 12 blood exposures listed in Table 1, cases were more likely than controls to have been exposed for eight exposures, controls were more likely than cases to have been exposed for one exposure and cases and controls were equally likely to have been exposed for three exposures. Each case had a greater number of types of exposures than his/her sibling control (except for one sibling pair who had the same number of exposures). Some blood exposures in the table actually included several potential types of exposures, such as dental surgery, which sometimes included anaesthetic injections. During the exposure period, no case or control had donated blood, received a blood transfusion or received an enema. All cases and controls had been shaved for hair cutting (on a regular basis and/or in traditional rituals in infancy).

All eight cases and all eight controls had had malaria previously according to their parents. All seven cases who had medical treatment for malaria had received injections and/or infusions during that treatment. However, only two of the five controls who had medical treatment for malaria received injections or infusions during that treatment. Thus, cases and controls did not differ much or at all in terms of malaria illness or medical treatment for malaria, but differed dramatically in terms of invasive medical treatment for malaria. Even after excluding malaria-related blood exposures, though, cases still had a significantly greater diversity of blood exposures (Table 1).

Cases and controls had similar reasons for most health-care-related blood exposures (Appendix). The reported reuse of sharps varied by the particular potential blood exposure (Appendix). Two of the eight families had kept syringes and injectable medicines at home as a general practice for many years. These syringes and medicines were shared freely within these nuclear families and other relatives in their extended families. Both families reported reusing the syringes, cleaning them by rinsing with hot water or boiling them, and then keeping them for several months to a year or more. Injectable medicines were kept until they expired. There was no evidence that the household syringes or injectable medicines were differentially shared with case and control children in these families.

The parents of four cases did not know how their children might have acquired HIV. The parents of the other four cases offered several different, and sometimes multiple, speculations about how their children became infected, including injections (n = 3), dental surgery (n = 2), contaminated barbering equipment (scissors, clippers) (n = 1), physical contact with blood (as from another's nosebleed or cut, n = 1), accidental needle prick (n = 1) and bewitching (n = 1).

DISCUSSION

We sought to estimate the extent and modes of horizontally transmitted HIV infection in Swazi and Kenyan children. In a national household probability sample, 20% of HIV-infected Swazi children aged 2–12 had HIV-negative mothers, indicating horizontally transmitted infections. In another study, HIV-infected Kenyan children aged 4–10 with HIV-negative mothers had more kinds of potential blood exposures than their uninfected siblings. Punctures related to health care for suspected malaria (phlebotomy, injection and infusion), injections while hospitalized and dental surgery (particularly by informal providers) were more common in infected children. Infected and uninfected children appeared similar in their reasons for seeking health care, and the perceived hygiene of invasive procedures varied by the particular exposure and setting. Some parents suspected blood exposures as the ways in which their children became infected.

The estimated proportion of Swazi children with horizontally-acquired HIV infection represents a lower bound because some HIV-positive children with HIV-positive mothers may have also acquired their infections horizontally.^2,20 Indeed, several blood exposures unrelated to possible symptoms or complications of HIV infection (such as vaccinations and injections for dental care) were associated with HIV infection in children of HIV-positive mothers in the Free State in 2004.¹ However, it is not unusual that all three HIV-infected Swazi children under age five who had HIV-negative mothers in the DHS had received all their recommended DPT, BCG, HBV and measles vaccinations, because 82% of all Swazi children under age 5 in the DHS had received all such vaccinations.¹⁵ Child sexual abuse is sometimes hypothesized as a potential mode of horizontal HIV transmission in children. Based on empirical estimates of the relevant factors, though, child sexual abuse is unlikely to account for a meaningful share of paediatric HIV infections in Swaziland (Appendix).

Some of the blood exposures that were more common in the HIV-infected Kenyan children than in their uninfected siblings have rarely been assessed in epidemiological research on HIV transmission in Africa, despite suspicions that they could transmit HIV and other bloodborne viruses. The practice of removing ‘abnormal’ or ‘false’ teeth (deciduous canine root buds) in infants to treat an acute illness or other problem is common in several areas of eastern Africa.^21,22 Observers and parents (including those in our study) have noted the potential of HIV transmission through unprotected, blood contaminated providers and cutting instrument reuse without sterilization or any cleaning whatsoever.^21,22 In addition, African public health officials have warned about the risk of HIV transmission when removing jiggers (chigoe or sand fleas) with reused, unsterile instruments.^23–25 To our knowledge, neither of these transmission modes have been highlighted in HIV prevention programmes designed or supported by international and foreign health aid agencies. In notable contrast, an indigenous Kenyan organization, the Ahadi Kenya Trust, has mounted a public education campaign about hygienic jigger removal.^23–25

Treatment for malaria may involve greater risk for HIV acquisition than treatment for many other common paediatric conditions because treatment often involves multiple injections and/or infusions for a single illness episode. Even when syringes for injecting patients tend not to be reused on different patients, multiple uses of an individual patient's syringe can contaminate medication vials that are then used on other patients. Similarly, unsterilized syringes reused in flushing intravenous catheters/cannulae or administering medicines into them also pose risk of transmitting bloodborne agents. Each of these practices has been implicated in nosocomial outbreaks of HIV, hepatitis C virus (HCV) and HBV.^26–30 Furthermore, phlebotomy was most frequently performed in the Kenyan children for malaria testing. Apart from potentially contaminated and reused phlebotomy needles/syringes and lancets, provider fingers and reused gloves, gauze, bandages and tourniquets may also be blood contaminated. Such potential blood exposures are common in many sub-Saharan African settings^31–36 and present additional risks of bloodborne virus transmission. Indeed, lack of glove use, blood-contaminated phlebotomy stations, and reused, blood-contaminated gloves and lancets were the most likely transmission modes in several nosocomial outbreaks of HBV and HCV.^37–42 If blood exposures during malaria treatment are involved with paediatric HIV transmission, then children who receive partially effective malaria vaccine⁴³ should also have lower HIV incidence than unvaccinated children.

The high rate of horizontal HIV infections in Swazi children and the associations of particular blood exposures with HIV infection in the Kenyan case-control study together strongly suggest that iatrogenic HIV transmission is not uncommon in these children. Our findings parallel substantial HIV prevalence rates in African virgins and adolescents and associations between potential blood exposures and HIV infection in these populations.^44–46 Moreover, our results dovetail with evidence that other bloodborne viruses, including HBV, HCV and HTLV-1, are transmitted horizontally in African children.^47–51

Both the Kenyan and Swazi studies we present have limitations. Given the informal nature of case referral, the Kenyan data cannot indicate the extent of horizontally transmitted HIV infection in Kenyan children. Despite the low likelihood of child sexual abuse accounting for significant horizontal transmission in Swaziland, parents in the Kenyan study might have underreported such abuse, either intentionally or because they were unaware. The Kenyan parents' reports may have been affected by recall bias in which parents may have been predisposed to recall blood exposures in their infected but not uninfected children. However, most parents did not suspect blood exposures as the mode of acquisition, and we also assessed sexual and breastfeeding exposures, which signalled the range of plausible transmission modes we were investigating. In the Kenyan study, we were unable to determine whether exposures preceded or followed HIV infection. For instance, it is possible that malaria treatment followed HIV infection for some cases. HIV infection might increase susceptibility to malaria infection.^18,19 This hypothesized relationship, though, might not necessarily be present in early and middle childhood in areas with endemic malaria, such as where the Kenyan children lived. We also did not assess some potential blood exposures in the Kenyan study. In addition, uninfected Kenyan sibling controls had 23% longer exposure periods than their infected siblings, which indicates that our results understate the degree to which cases' incidence of blood exposures surpassed that of controls. The generalizability of the Swazi results is limited by the large percentage of Swazi children (approximately one-third between age two and 14) who did not live with either parent (due to orphanhood or other reasons)¹⁵ and thus were not included in any of our analyses. Finally, both the Swazi and Kenyan studies lacked objective genetic confirmation of biological maternity, although the documentation of biological maternity in the Kenyan study was sound.

To determine the precise routes of horizontal transmission in sub-Saharan African children, stronger research designs are required. These designs would include comprehensive assessment of blood, sexual and breastfeeding exposures in cases and controls, tracing their contacts involved in such exposures, and comparison of presumed sources' and child cases' HIV DNA sequences.^52,53 Such investigations could be a standard response to paediatric cases of horizontally-transmitted HIV infection that would be identified through formal surveillance systems relying on standardized protocols for testing mothers of all children diagnosed with HIV infection. To our knowledge, no such surveillance systems have been implemented anywhere in sub-Saharan Africa, despite the urgent need. Similarly, HIV prevention campaigns in all populations should be made whole by highlighting the full range of bloodborne risks and emphasizing practical ways to avoid exposure.^54,55

Footnotes

Acknowledgements

We thank Margaret Natocho, Peter Opande, Sabina Magero and Mildred Boyani for their assistance with data collection. A de-identified data-set for the Kenyan case-control study is available at: http://dvn.iq.harvard.edu/dvn/dv/SHA.

Appendix

References

Shisana

, Connolly

, Rehle

, Mehtar

, Dana

. HIV risk exposure among South African children in public health facilities. AIDS Care 2008;20:755–63

Gisselquist

, Potterat

, Brody

. HIV transmission during paediatric health care in sub-Saharan Africa–risks and evidence. S Afr Med J 2004;94:109–16

Hiemstra

, Rabie

, Schaaf

, Unexplained HIV-1 infection in children – documenting cases and assessing for possible risk factors. S Afr Med J 2004;94:188–93

Omotade

, Olaleye

, Saliu

, Odaibo

, Adeyemo

. Human immunodeficiency seropositivity among mother–child pairs in South West Nigeria: a community-based survey. West Afr J Med 2001;20:232–6

Adejuyigbe

, Durosinmi

, Onyia

, Adeodu

. Blood transfusion related paediatric HIV/AIDS in Ile-Ife, Nigeria. AIDS Care 2003;15:329–35

Biraro

, Morison

, Nakiyingi-Miiro

, Whitworth

, Grosskurth

. The role of vertical transmission and health care-related factors in HIV infection of children: a community study in rural Uganda. J Acquir Immune Defic Syndr 2007;44:222–8

Kengeya-Kayondo

, Malamba

, Nunn

, Seeley

, Ssali

, Mulder

. Human immunodeficiency virus (HIV-1) seropositivity among children in a rural population of south-west Uganda: probable routes of exposure. Ann Trop Paediatr 1995;15:115–20

Uganda Ministry of Health, ORC Macro. Uganda HIV/AIDS Sero-behavioural Survey 2004–2005. Calverton, MD, USA: Ministry of Health and ORC Macro, 2006

Onankpa

, Airede

, Paul

, Dorcas

. Pattern of pediatric HIV/AIDS: a five-year experience in a tertiary hospital. J Natl Med Assoc 2008;100:821–5

10.

Etokidem

. HIV/AIDS Transmission Through Female Genital Cutting: A Case Report. Abstract No. D10677, 15th International Conference on AIDS, Bangkok, 11–16 July, 2004

11.

Addo-Yobo

, Lovel

, Aidoo

. Malnutrition and malaria gave him anaemia, then the blood transfusion gave him HIV. Trop Doct 1992;22:123

12.

Oniyangi

, Awani

, Iregbu

. The pattern of paediatric HIV/AIDS as seen at the National Hospital Abuja Nigeria. Niger J Clin Pract 2006;9:153–8

13.

Rehle

, Shisana

, Pillay

, Zuma

, Puren

, Parker

. National HIV incidence measures–new insights into the South African epidemic. S Afr Med J 2007;97:194–9

14.

Mann

, Francis

, Davachi

, Risk factors for human immunodeficiency virus seropositivity among children 1–24 months old in Kinshasa, Zaire. Lancet 1986;2:654–7

15.

Central Statistical Office (Swaziland), Macro International. Swaziland Demographic and Health Survey 2006–07. Mbane, Swaziland: Central Statistical Office and Macro International Inc, 2008

16.

Central Statistical Office (Swaziland), Macro International. Swaziland Demographic and Health Survey 2006–07, Preliminary Report. Calverton, MD, USA: Measure DHS, Macro International, 2007

17.

U.S. Census Bureau. International Data Base (IDB). Available at: http://www.census.gov/ipc/www/idb/ (last checked 29 January 2009)

18.

Laufer

, Plowe

. The interaction between HIV and malaria in Africa. Curr Infect Dis Rep 2007;9:47–54

19.

Slutsker

, Marston

. HIV and malaria: interactions and implications. Curr Opin Infect Dis 2007;20:3–10

20.

Brewer

, Gisselquist

, Brody

, Potterat

. Investigating iatrogenic HIV transmission in Ugandan children. J Acquir Immune Defic Syndr 2007;45:253–4

21.

Miles

, Ololo

. Traditional surgeons in sub-Saharan Africa: images from south Sudan. Int J STD AIDS 2003;14:505–8

22.

Accorsi

, Fabiani

, Ferrarese

, Iriso

, Lukwiya

, Declich

. The burden of traditional practices, ebino and tea-tea, on child health in Northern Uganda. Soc Sci Med 2003;57:2183–91

23.

Kumba

. HIV twist to jigger menace. Daily Nation (Nairobi), 18 October 2008

24.

Nabusoba

. Uganda: jiggers – a neglected public health concern. New Vision (Kampala) , 20 April 2008

25.

Wangui

. Kenya: rural dwellers face brunt of jiggers. Africa News. Available at: http://www.africanews.com/site/list_messages/20928 , 8 October 2008 (last checked 31 October 2008)

26.

Katzenstein

, Jorgensen

, Permin

, Nosocomial HIV-transmission in an outpatient clinic detected by epidemiological and phylogenetic analyses. AIDS 1999;13:1737–44

27.

Lagging

, Aneman

, Nenonen

, Nosocomial transmission of HCV in a cardiology ward during the window phase of infection: an epidemiological and molecular investigation. Scand J Infect Dis 2002;34:580–2

28.

Krause

, Trepka

, Whisenhunt

, Nosocomial transmission of hepatitis C virus associated with the use of multidose saline vials. Infect Control Hosp Epidemiol 2003;24:122–7

29.

Massari

, Petrosillo

, Ippolito

, Transmission of hepatitis C virus in a gynecological surgery setting. J Clin Microbiol 2001;39:2860–3

30.

Thompson

, Perz

, Moorman

, Holmberg

. Nonhospital health care-associated hepatitis B and C virus transmission: United States, 1998–2008. Ann Intern Med 2009;150:33–9

31.

Cronin

, Quansah

, Larson

. Obstetric infection control in a developing country. J Obstet Gynecol Neonatal Nurs 1993;22:137–44

32.

Bernath

. Tourniquets in Phlebotomy. Evidence Centre Evidence Report Series 2002: Intervention, Monash University; 2002. Available at: http://www.med.monash.edu.au/healthservices/cce (last checked 6 January 2009)

33.

Golder

, Chan

, O'Shea

, Corbett

, Chrystie

, French

. Potential risk of cross-infection during peripheral-venous access by contamination of tourniquets. Lancet 2000;355:44

34.

Rourke

, Bates

, Read

. Poor hospital infection control practice in venepuncture and use of tourniquets. J Hosp Infect 2001;49:59–61

35.

Meili

. A Message from Mozambique: the Impact of Conflict on Health Infrastructure and, Consequently, Kids. Paper presented at International Conference on Children and War: Impact, Edmonton, Canada, 1–4 April 2004. Available at: www.medicine.usask.ca/dean/cc_social_acc/mtl/pdfs/Message%20from%20Mozambique.pdf (last checked 13 January 2009)

36.

The University of East Africa: the excitement of learning. Scope 1996;(Autumn):34–5. Available at: www.llu.edu/news/scope/aut96/34.pdf (last checked 13 January 2009)

37.

Centers for Disease Control and Prevention. Nosocomial hepatitis B virus infection associated with reusable fingerstick blood sampling devices – Ohio and New York City, 1996. Morb Mortal Wkly Rep 1997;46:217–21

38.

Polish

, Shapiro

, Bauer

, Nosocomial transmission of hepatitis B virus associated with the use of a spring-loaded finger-stick device. N Engl J Med 1992;326:721–5

39.

Dibenedetto

, Miraglia

, Ippolito

, Reduction in the incidence of infection by hepatitis C virus in children with acute lymphoblastic leukemia after suspension of sampling from the finger. Pediatr Infect Dis J 1996;15:265–6

40.

Dreesman

, Baillot

, Hamschmidt

, Monazahian

, Wend

, Gerlich

. Outbreak of hepatitis B in a nursing home associated with capillary blood sampling. Epidemiol Infect 2006;134:1102–13

41.

Quale

, Landman

, Wallace

, Atwood

, Ditore

, Fruchter

. Deja vu: nosocomial hepatitis B virus transmission and fingerstick monitoring. Am J Med 1998;105:296–301

42.

Niu

, Alter

, Kristensen

, Margolis

. Outbreak of hemodialysis-associated non-A, non-B hepatitis and correlation with antibody to hepatitis C virus. Am J Kidney Dis 1992;19:345–52

43.

Bejon

, Lusingu

, Olotu

, Efficacy of RTS,S/AS01E vaccine against malaria in children 5 to 17 months of age. N Engl J Med 2008;359:2521–32

44.

Brewer

, Potterat

, Roberts

, Brody

. Male and female circumcision associated with prevalent HIV infection in adolescents in Kenya, Lesotho, and Tanzania. Ann Epidemiol 2007;17:217–26

45.

Brewer

, Potterat

, Muth

, Brody

. Converging evidence suggests nonsexual HIV transmission among adolescents in sub-Saharan Africa. J Adolesc Health 2007;40:290–1

46.

Gavin

, Galavotti

, Dube

, Factors associated with HIV infection in adolescent females in Zimbabwe. J Adolesc Health 2006;39:596.e11–18

47.

Abdool Karim

, Coovadia

, Windsor

, Thejpal

, van den Ende

, Fouche

. The prevalence and transmission of hepatitis B virus infection in urban, rural and institutionalized black children of Natal/KwaZulu, South Africa. Int J Epidemiol 1988;17:168–73

48.

Vall Mayans

, Hall

, Inskip

. Do bedbugs transmit hepatitis B? Lancet 1994;344:962

49.

Ngatchu

, Stroffolini

, Rapicetta

, Chionne

, Lantum

, Chiaramonte

. Seroprevalence of anti-HCV in an urban child population: a pilot survey in a developing area, Cameroon. J Trop Med Hyg 1992;95:57–61

50.

Biggar

, Ortiz-Conde

, Bagni

, Hepatitis C virus genotype 4 in Ugandan children and their mothers. Emerg Infect Dis 2006;12:1440–3

51.

Olaleye

, Omotade

, Sheng

, Adeyemo

, Odaibo

. Human T-cell lymphotropic virus types I and II infections in mother–child pairs in Nigeria. J Trop Pediatr 1999;45:66–70

52.

Brewer

, Rothenberg

, Potterat

, Brody

, Gisselquist

. HIV epidemiology in sub-Saharan Africa: rich in conjecture, poor in data. Int J STD AIDS 2004;15:63–5

53.

Brewer

, Hagan

, Sullivan

, Social structural and behavioral underpinnings of hyperendemic Hepatitis C virus transmission in drug injectors. J Infect Dis 2006;194:764–72

54.

Gisselquist

, Friedman

, Potterat

, Minkin

, Brody

. Four policies to reduce HIV transmission through unsterile health care. Int J STD AIDS 2003;14:717–22

55.

Correa

, Gisselquist

, Gore

. Blood-borne HIV: Risks and Prevention. Chennai, India: Orient Longman Private Limited, 2008

56.

Reza

, Breiding

, Blanton

, Violence Against Children in Swaziland. UNICEF. Available at: http://www.unicef.org/swaziland/sz_publications_2007violenceagainstchildren.pdf , 2007 (last checked 21 December 2008)