Zimbabwe: what will it take to achieve virtual elimination of infant HIV infections?

Efforts to achieve "virtual elimination" of new HIV infections in infants by 2015 could prove highly challenging, according to computer modelling of the situation in Zimbabwe, one of the most seriously HIV-affected countries in southern Africa.

Using a validated computer simulation model of a cohort of HIV-infected, pregnant or breastfeeding women in Zimbabwe, researchers found that even with a 95% uptake of the World Health Organization’s (WHO) 2010 prevention of mother-to-child (PMTCT) guidelines, projected transmission risks (6.1 to 7.7%) would still exceed those set for WHO’s goal of  “virtual elimination” of paediatric HIV (defined as a mother-to-child-transmission risk of less than 5%).

This analysis, published in the 10 January edition of PLoS Medicine, looked at the effectiveness of three PMTCT treatment regimens:

• single-dose nevirapine (sdNVP)

• WHO's Option A regimen: zidovudine during pregnancy and sdNVP at delivery and infant nevirapine throughout breastfeeding in women with advanced illness

• the Option B regimen: triple ART during pregnancy and breastfeeding and lifelong ART for women with advanced illness.

These were studied at four different levels of uptake of PMTCT services (36, 56, 80 and 95%, corresponding to reported national 2008, 2009 rates and WHO target and optimal rates, respectively).

The authors found that increased uptake of PMTCT services played a significant role in reducing MTCT, even in times of economic hyperinflation; uptake from 2008 to 2009 increased from 36 to 56%, with a corresponding decrease in MTCT from 20.3 to 18%.

Replacing the 2009 national programme of sdNVP with Option A (at the corresponding reported uptake of 56%), was projected to reduce the estimated transmission risk from 18% to 14.4%, or 13.4% with Option B, they add. Both options are more effective than increasing sdNVP coverage to 80% (which has a MTCT risk of 15.4%).

The planned implementation of Option A in Zimbabwe, like many other sub-Saharan African countries where prolonged breastfeeding is common, could significantly reduce infant HIV infection, compared to the 2009 national programme of sdNVP.

In the United States and Europe, the availability of antiretroviral therapy during pregnancy and avoidance of breastfeeding among HIV-infected women has led to the near elimination of paediatric HIV.

HIV can be passed from the mother to the child during pregnancy, labour and delivery, or breastfeeding. Without treatment, approximately 30% of infants will be infected during pregnancy or delivery, and an estimated 5 to 20% during breastfeeding.

In settings where breastfeeding is the norm and replacement feeding is not an option, recent studies have shown that the provision of maternal ARV during breastfeeding can reduce MTCT to between 1 and 5%.

As a consequence, WHO, among others, made “virtual elimination” of paediatric HIV a new target and accordingly released new guidelines in 2010 comprising options A and B.

Yet many women are lost to care at any point along what is known as the PMTCT ‘cascade’:

• receiving antenatal care

• having an HIV test and getting the results

• being assessed for ART eligibility

• having a CD4 cell count test and timely receipt of results

• availability of ART

• help with adherence

• receiving postnatal care. 

Only an estimated 53% of pregnant women got any kind of treatment for PMTCT in 2009, resulting in around 400,000 infants newly infected with HIV; over 90% were in sub-Saharan Africa.

Given the difficulties of enrolling and keeping women in care at any step of the PMTCT cascade, the effectiveness of Options A or B is unknown.

So the authors chose to expand a simulated model of MTCT to include all the steps, projecting the level of PMTCT uptake, the drug regimens, and length of breastfeeding necessary, to reach “virtual elimination” in Zimbabwe.

The model simulated a cohort of two populations of HIV-infected pregnant and breastfeeding women, with a mean age of 24, a mean CD4 cell count of 451cells/mm³, and a breastfeeding period of 12 months. The cohort was followed from first antenatal visit to two years after giving birth.

Cohort 1 included women already HIV-infected at their first antenatal visit. Cohort 2 included all women getting pregnant each year in Zimbabwe (392,460) with an HIV prevalence of 16% at the first antenatal visit and HIV incidence of 1% during late pregnancy and breastfeeding. Cohort 1 was nested within Cohort 2. Cohort 1 was analysed to project MTCT rates and Cohort 2 to project the proportion and number of infants to become HIV-infected in an annual birth cohort.

The primary outcome was MTCT risk at weaning. The authors looked at improvements in two measures of PMTCT care: more effective regimens and improved uptake of the PMTCT 'cascade'.

The authors note their findings are consistent with previous analyses (using different modelling methodologies) that suggest that “virtual elimination” will require massive scale-up of PMTCT services.

The authors stress that, regardless of whether sdNVP was replaced with Option A or B, it is the level of PMTCT uptake that has the greatest effect on MTCT rates.

They note that, once 56% uptake is reached with Option A, increasing this to 80% results in an MTCT risk of 10.5%, with better outcomes than if Option A were replaced with Option B at current levels of 56% (an MTCT risk of 13.4%).

This analysis depends on the effectiveness of each ART regimen, note the authors, adding that Option B may not be better than Option A for PMTCT; no randomised comparisons have been reported. Studies suggest that both options are equally efficacious in women with CD4 cell counts over 350 cells/mm³, in line with WHO’s strong recommendation for both options.

They cite the anticipated results of the IMPAACT/PROMISE study, a multinational clinical trial comparing these regimens in women with CD4 cell counts over 350 cells/mm³ which will, importantly, inform policy decisions around Options A and B.

As noted, increased uptake of PMTCT services reduced projected MTCT risk. This happens, the authors add, by improving the proportion of ART-eligible women who get ART, including the greater availability of CD4 testing and timely receipt of results, or by providing triple-drug ART.

While shorter breastfeeding periods can contribute to reduced MTCT, the absence of clean water and availability of infant formula make this impractical in many resource-poor settings.

The specific step in the cascade where uptake varied did not affect MTCT risk at 4 to 6 weeks. However, the proportion of mother-infant pairs linked in to postnatal care had a considerable effect on twelve-month MTCT risk.

The authors suggest promotion of high levels of access to postnatal care and adherence to ARV during breastfeeding is critical.

Limitations of the study included computer model simplification of complex biological and operational processes; data from multiple sources was combined.

To meet this challenging and critical goal of “virtual elimination” in Zimbabwe, or in other resource-poor settings with a high HIV prevalence and prolonged breastfeeding, a national programme using either Option A or B must be implemented, together with strategies to improve access to PMTCT services (to close to 100%), to retain women in care, and to support medication adherence throughout pregnancy and breastfeeding, the authors conclude.

Ciaranello AL et al. What will it take to eliminate pediatric HIV? Reaching WHO target rates of mother-to-child HIV transmission in Zimbabwe: a model-based analysis. PLoS Medicine, Volume 9, Issue 1, e1001156, January 2012.