Subscribe to:
POZ magazine
Join POZ: Facebook MySpace Twitter Pinterest
Tumblr Google+ Flickr MySpace
POZ Personals
Sign In / Join

Transmission of HIV From Infants To Their Mothers


This title may surprise some.   In a paper (abstract, below) from a group at CDC I learned yet another HIV/AIDS related acronym.  It's CBWT, Child-to-Breastfeeding-Woman Transmission.  


 There have been several reports over many years of HIV infected infants born to mothers who were HIV uninfected.   These infections were noted as early as the late 1980s in the former Soviet Union, in Libya in 1998, in Kyrgyzstan, Kazakhstan, Romania as well as in Africa.   In every instance except in Africa, there cases were investigated with varying degrees of thoroughness.   The sources of infection were invariably associated with contaminated blood,   either from transfusions, or from procedures in unsafe healthcare settings, where for example sterilization of instruments is inadequate, or injection equipment is reused.

The infections noted in infants that were investigated occurred as outbreaks and all were determined to be nosocomial.    Although infected infants born to uninfected mothers have been noted in Africa, remarkably, it appears that none have been investigated.

 It will probably remain for a future historian to understand why cases of HIV infection in infants, horizontally rather than vertically transmitted, have yet to be investigated in Africa.



 In those non-African outbreaks that were investigated transmission occurred through unsafe medical care, so what do we know of the safety of health care facilities in Africa ?

Unfortunately unsafe health care remains a problem in many facilities in high prevalence areas in Africa.



Taking Kenya as an example, Simon Collery has written in one of my blogs:

Where does Kenya fit into this picture? As UNAIDS admit, there's not much data. But there is a document called the Service Provision Assessment  which looks at conditions in various kinds of health facility, such as hospitals, clinics and pharmacies.


A few samples from this document may suffice to illustrate Kenya's capacity to prevent HIV transmission through unsafe injections and other healthcare practices: Between 10 and 15% of facilities don't have adequate supplies of needles, syringes or latex gloves; between 55 and 70% don't have running water or soap; many don't have facilities for disposing of contaminated equipment or supplies of disinfectant; less than half have guidelines for infection prevention and less than 10% have guidelines for sterilization. 

Although this document dates from 2004, we don't know if there has been any change. 

Kenya HIV2.jpg

As the title of this post indicates, infants infected either vertically or through exposure to contaminated blood are able to transmit HIV to seronegative women who breast feed them.


Mother to child transmission is the leading cause of HIV infection in infants. Some of these infected infants will be orphans and so place seronegative women who breastfeed them at risk.  Wet-nursing is the complete nursing of another woman's infant and still occurs as does cross-nursing which is the nursing of another infant by a woman  while still nursing her own child.  Estimates of the prevalence of these practices vary by region and the overall prevalence is not known.

Worldwide the greatest risk for CBWT is carried by seronegative mothers whose infants become infected through exposure to contaminated blood.  Rates of CBWT were as high as 40-60% among mothers  breastfeeding infants who became infected after birth. (noted in the paper referenced below).

This report on CBWT highlights the importance of unsafe health care facilities in the transmission of HIV.   Of course HIV is not the only pathogen that can be transmitted in such settings.  

Why so few resources have been devoted to the improvement of health care facilities in developing nations is puzzling.  Could it be that like the provision of clean water and sanitation, improving health care facilities is not something that can generate much profit?

Perhaps it will be left to HIV activists who have successfully drawn public attention to other neglected issues, to alert funders and policy makers to the dangerous condition of many healthcare facilities in the developing world.

The benefits of improving infection control in these facilities extend far beyond effects in HIV transmission.

A group of individuals have been trying to bring attention to this issue for many years and I do recommend looking at the website they have created to directly alert people in Africa to dangers in health care facilities with no or poor infection control procedures.


I realize I neglected one point,

 Exposure to saliva is not considered to be a great risk for HIV transmission, so why is it happening in the case of CBWT?

The authors suggest maternal skin fissures or infant stomatitis (inflammation of the lining of the mouth).  I'm not aware of research that would explain such transmissions in infancy. I hope it will be undertaken. 





A Review of Evidence for Transmission of Human Immunodeficiency Virus from Children to Breastfeeding Women and Implications for Prevention.


Kirsten M Little, Peter Kilmarx, Allan Taylor, Charles Rose, Emilia Rivadeneira. And Steven Nesheim.

The Pediatric Infectious Disease Journal Publish ahead of print.



Background: Child-to-Breastfeeding-Woman Transmission (CBWT) of HIV occurs when an HIV-infected infant transmits the virus to an HIV-uninfected woman through breastfeeding. Transmission likely occurs as a result of breastfeeding contact during a period of epithelial disruption, such as maternal skin fissures and/or infant stomatitis. Despite extensive epidemiologic and phylogenetic evidence, however, CBWT of HIV continues to be overlooked.

Objective: This paper summarizes the available evidence for CBWT from nosocomial outbreaks, during which nosocomially HIV-infected infants transmitted the virus to their mothers through breastfeeding. This paper also explores the CBWT risk associated with HIV-infected orphans and their female caretakers, and the lack of guidance regarding CBWT prevention in infant feeding recommendations.

Methods: We searched online databases including PubMed and ScienceDirect for English language articles published from January 1975 to January 2011 using the search terms "HIV", "perinatal", "child-to-mother", and "breastfeeding". The citations from all selected articles were reviewed for additional studies.

Results: We identified five studies documenting cases of CBWT. Two studies contained data on the number of HIV-infected women, as well as the proportion breastfeeding. Rates of CBWT ranged from 40 - 60% among women reporting breastfeeding after their infants were infected.

Conclusions: Poor infection control practices, especially in areas of high HIV prevalence, have resulted in pediatric HIV infections and put breastfeeding women at risk for CBWT. Current infant feeding guidelines and HIV prevention messages do not address CBWT, and fail to provide strategies to help women reduce their risk of acquiring HIV during breastfeeding.



Show Comment(s)

Comments on Joseph Sonnabend, MD's blog entry "Transmission of HIV From Infants To Their Mothers"

"Although infected infants born to uninfected mothers have been noted in Africa, remarkably, it appears that none have been investigated."

Full text:

AIDS Care. 2008 Aug;20(7):755-63.

HIV risk exposure among South African children in public health facilities.

Shisana O, Connolly C, Rehle TM, Mehtar S, Dana P.

Human Science Research Council, Cape Town, South Africa.


The study investigates the risk exposure to HIV infection among South African children aged 2-9 years served by public health services. Together with their biological mothers, 3471 children and were recruited from inpatient and outpatient children in the Free State Province. Blood samples were taken by professional nurses and a history taken of exposure factors associated with HIV transmission. DNA testing was used to confirm biological maternity where the child was HIV-positive and the mother HIV-negative. Mother-child pairs were stratified by mother's HIV status. Exposure factors related to the child's HIV status were examined in each stratum using a chi-square test. Independent factors were then included in a multiple logistic regression model. Having an HIV-positive mother was strongly related to HIV infection in children (OR: 310; 95%CI: 148-781). However, seven HIV-positive children had HIV-negative mothers. Transmission in this group was significantly associated with breastfeeding by a non-biological mother (OR: 437; 95%CI: 53-5020), being fed with expressed breast milk from a milk room (OR: 37.6; 95%CI: 6.2-259.0), dental injection history (OR: 31.5; 95%CI: 4.5-189.4) and visits to a dentist (OR: 26.9; 95%CI: 4.4-283.5). Although mother-to-child-transmission is shown to be the primary mode of HIV transmission in South African children, the few HIV-positive children infected by other modes of transmission suggest a potential risk of non-vertical HIV infections. These infections can be prevented through education and improved infection-control procedures.

The data reported here are consistent with this newly reported case that was confirmed by sequencing:

- Your erstwhile co-author David Gisselquist maintains a list of previously published papers pertaining to South Africa, but he does not include the one above:

Thanks very much for bringing this study to attention, I'm sorry it was missed. . The importance of improved infection control is noted in the abstract which is what was emphasized in the post.

Any idea as to why it is not included on David Gisselquist's "Don't Get Stuck" website, which is now referenced continually on the POZ Magazine website? It is directly relevant to his claims and was published in 2008 (with the data presented prior to that). Do you think it is doing a service to POZ readers to refer them to a site which takes a Duesbergian approach to the evidence i.e. trumpeting any study that can be construed as supporting the theory of Gisselquist et al while pretending the evidence that refutes it does not exist?

I recommend reading the full text of this paper.

"Among mother-child pairs with HIV-negative mothers, HIV infection was not associated with age group; prior history of hospital admission; history of blood transfusion; injections; immunisations; visiting a traditional healer; or scarification."

Hi Joseph and Richard,

Good to have more attention to HIV-positive children with HIV-negative mothers!

Richard, the study you cite (Shisana et al, AIDS CARE 2008, pp755ff) reports four significant risks for HIV in children with HIV-negative mothers (adjusted OR > 1 and p < 0.001): (a) ever visited a dentist; (b) injection at a dental care facility; (c) breastfed by a non-biological mother; and (d) fed with breastmilk from a milkroom (in a hospital). This study adds to evidence that skin-piercing procedures can be dangerous for children.

Many other studies find HIV in African children with HIV-negative mothers along with evidence linking infections to blood risks (see

For example, a recent random sample national survey in Mozambique identified 63 HIV-positive children aged 0-11 years, of which 29% (18/58) had HIV-negative mothers; 5 of 63 mothers were not tested. In children with HIV-negative mothers and in youth to age 19 years, HIV was significantly associated with male circumcision and scarification (see: and

(continued in next comment)

(continued from previous comment)
No government in sub-Saharan Africa has investigated any of thousands of identified HIV-positive children with HIV-negative mothers by tracing and testing other children who visited the same health facilities (as was done in Russia in 1988-89 to find >260 children infected; in Libya in 1998-99 to find >400 children infected; and in other countries outside sub-Saharan Africa).
There is nothing Duesbergian about advocating safe health care for Africans, and extending to Africans the same warnings about unsafe health care that UNAIDS gives to UN employees (see page 9 in this link:
In more than 20 published papers I report that HIV from sex and skin-piercing procedures infects and kills far too many Africans. Let's work together to stop it.

What was the explanation for the omission of the paper from your website David? I don't see it in your comments.

As I note above, the study you brought into this discussion reports that injections and dental care are related to HIV in kids aged 2-7 years with HIV-negative mothers. Another paper from the same study documents reuse of bloody instruments in dental care (see:
There is so much evidence of suspected nosocomial HIV infections in kids in Africa that I don't try to list it all in dontgetstuck. In the same way if I wrote about rainfall in Africa, I couldn't list all the thunderstorms. Too many.

It's relevant to point out some details of the peer-reviewed Shisana paper. It identifies seven HIV-positive children with HIV-negative mothers, compared with 477 with HIV-positive mothers:

"Of the 1014 children with HIV-positive mothers, 477 (47%) were HIV-positive. As anticipated, the factor most strongly associated with HIV infection among children was having a mother who was HIV-positive (OR: 310; 95%CI: 148-781). However, of the 2457 children with HIV-negative mothers, seven (0.28%) were HIV-positive."

The paper lists the odds ratios for the factors significantly associated with seropositivity in these seven children, in order of magnitude. Even in his internet comment above, David Gisselquist is unable to stop himself reversing this order to place the putative injection risk first, and omitting reference to the size of the odds ratios involved. The desire to obfuscate seems reflexive. But as can be seen from the abstract above, the odds ratio associated with breast feeding by a non-biological mother was massively higher than the other three:

"Transmission in this group was significantly associated with breastfeeding by a non-biological mother (OR: 437; 95%CI: 53-5020)"

Which is consistent with the DNA-confirmed case report in the Aug 18 issue of the Lancet:

Furthermore, I don't think anyone would deny the importance of safe dental practice, but the paper also shows that many children in this study had been exposed to dentists and dental injections without acquiring infection:

"Exposure to dentists and dental injections was slightly higher with 210 (8.7%) of children visiting a dentist and 60 (2.4%) receiving dental injections."

Not only does David Gisselquist not mention this paper on his "Don't Get Stuck" website, he also--as far as I can tell--omits any reference to it in his updated online book "Points to Consider" despite the fact that it discusses the issue and cites other work by the same author (and includes citations up to 2009):

And that's exactly the type of shameful disingenuousness that renders you an untrustworthy and non-credible source. You've been well aware of these data for years, your omission of any reference to them is deliberate.

The dense, technical nature of your oeuvre often renders what you do opaque to most readers, so what you did in your comment above is usefully illustrative because it is so transparent. Even the abstract of the paper , posted here, lists the odds ratios for the *associations* with HIV infection in the *seven* children identified. I don't think it requires expertise in odds ratios to appreciate that the first one--breastfeeding by a non-biological mother, with an odds ratio of 437--is a massively stronger association than the other three, where the ratios range from 26-37.

How do you handle this in your comment? In order to avoid addressing the huge differences in the odds ratios, you write that the paper "reports four significant risks for HIV in children with HIV-negative mothers (adjusted OR > 1." The sleight of hand of grouping them only based on whether the odds ratios were greater than 1 then allows you to list them in your preferred order, rather than in order of the magnitude of the associations (as the authors did). And you performed this manipulation just for the purposes of a comment on a blog.

The figures also aren't adjusted, because there were two few children with HIV-negative mothers:

"As only seven of the children with an HIV-negative mother were HIV-positive, it was not possible to adjust the odds ratios for potential confounders among children with HIV-negative mothers."

Your concern about HIV transmission from surrogate breastfeeders to babies is well taken. I have revised dontgetstuck to include that message:
But we still have the question: How important is surrogate breastfeeding as a risk for HIV in children? The Shisana study you brought into this discussion ( reports that surrogate breastfeeding is a bigger risk for a child to be HIV-positive than having an HIV-positive mother. The study team could have tested this surprising claim by tracing and testing surrogate breastfeeders as well as other children visiting suspected clinics and hospitals.
Tracing infections to find their source is decades overdue in Africa. I invite you to join me and others in calls for tracing to find the risks that infect so many children and adults.

I tried to respond to this latest appalling misrepresentation when it was first posted but was stymied by POZ's blog software. David Gisselquist writes

'The Shisana study you brought into this discussion reports that surrogate breastfeeding is a bigger risk for a child to be HIV-positive than having an HIV-positive mother. The study team could have tested this surprising claim by tracing and testing surrogate breastfeeders as well as other children visiting suspected clinics and hospitals.'

This again offers insight into Gisselquist's mode of operation. The claim in the first sentence is NOT TRUE. It is based on either his inability to understand the paper, or his desperation to misrepresent it in order to pretend it somehow supports his position. Worse, in the next sentence he tries to offload his own misrepresentation onto the paper's authors, writing that it is THEIR 'surprising claim.' THEY MADE NO SUCH CLAIM. These misleading sleights of hand are typical of his writings.

If you read the paper, the data unequivocally show that the 'bigger risk for a child to be HIV-positive' is having an HIV-positive mother. Out of 1014 children born to HIV-positive mothers, 477 (47%) were HIV-positive. Out of 2457 children born to HIV-negative mothers, 7 (0.28%) of were HIV-positive. 477 is much bigger than 7, and 47% is much bigger than 0.28%.

Furthermore, the authors report the odds ratio for the association between being HIV positive and 'having been breastfed by a non-biological mother' among the children with HIV-positive mothers - it is 5 (95%CI 2.3-10.8), far lower than the odds ratio for the association with having an HIV-positive mother which is 310 (95%CI 148-781).

The seven HIV-positive children with HIV-negative mothers cannot have an odds ratio calculated for the association between infection and having an HIV-positive mother because they don’t have HIV-positive mothers. In this group, the odds ratio for the association between breastfeeding by a non-biological mother and being HIV-positive is 437 (95%CI 53-5020). Gisselquist is trying to make something of the fact that this number is numerically higher than 310 in the other group when this is a totally specious comparison. It reveals a complete misunderstanding of--and worse, willingness to abuse--the use of odds ratios. In his own blog post on the subject, he makes the common mistake of confusing odds ratios with relative risk by stating 'According to these data, breastfeeding with another women increased the child’s risk to be HIV-positive by 437 times.'

Gisselquist also fails to note the key difference between this paper and the other reports he tries to cite in support of his views - it involves DNA analyses of biological mother-child relationships.

It is both mystifying and depressing to me that Joe Sonnabend, who I'm sure understands these issues far better than I do, is collaborating with this person.

Hi Richard and others following this exchange:
Let’s not get distracted. The point that Joseph makes in the blog post that started these comments is that unexpected HIV infections in children in sub-Saharan Africa (with HIV-negative mothers) have not been adequately investigated. An adequate investigation involves tracing and testing other children attending suspected health care facilities. In Russia, Romania, Libya, and other countries, governments have investigated unexpected HIV infections in children in that way – eg, an investigation in Russia in 1988-89 found more than 260 children infected through hospitals. Similarly, in the US, investigations to find bloodborne infections from health care have been common over the past decade and have found lots of HBV and HCV infections.
When a child is found to be HIV-positive with an HIV-negative mother, we can speculate that the child might have gotten HIV from some source other than health care – eg, surrogate breastfeeding. But if it’s possible the HIV came from health care, the government should trace and test other children attending suspected facilities. Did the facilities infect other children? Is health care safe? Speculation about other risks does not answer these questions. Investigations are required.
Finally, let’s review the curiously high risk to transmit HIV through surrogate breastfeeding according to data reported in Shisana et al (see Table 3 in: The article reports: 15 children with HIV negative mothers breastfed with other women; 4 of these 15 children were HIV-positive; and an odds ratio = 437 to get HIV associated with surrogate breastfeeding. These data suggest that all 4 HIV-positive children with HIV-negative mothers who breastfed with other women got HIV from surrogate breastfeeding. That’s a curious result: Based on HIV prevalence among women in the community, only 5 (1/3) of 15 surrogate breastfeeders would have been HIV-positive, so the efficiency of HIV transmission from surrogate breasfeeders to children would be circa 80% (5 infected women infect 4 children). That’s much greater than the 35% rate of transmission from HIV-positive mothers to babies delivered and breastfed for 2 years without antivirals. The study team did not investigate and clarify their reported high risk to transmit through surrogate breastfeeding by tracing and testing surrogate breastfeeders and sequencing viruses.

Leave a comment



My Favorite Links

Subscribe to Blog

About this Entry

This page contains a single entry by Joseph Sonnabend published on June 29, 2012 3:25 PM.

Have we ignored a very simple procedure that could significantly reduce the risk of sexual transmission of HIV to men from women? was the previous entry in this blog.

How to have sex in an epidemic: 30th anniversary is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.


The opinions expressed by the bloggers and by people providing comments are theirs alone. They do not necessarily reflect the opinions of Smart + Strong and/or its employees.

Smart + Strong is not responsible for the accuracy of any of the information contained in the blogs or within any comments posted to the blogs.

© 2015 Smart + Strong. All Rights Reserved. Terms of use and Your privacy