Prevalence and correlates of HIV among children born to HIV-negative mothers in Zambia: a secondary analysis of the 2024 Zambia Demographic and Health Survey

This secondary analysis of the 2024 Zambia Demographic and Health Survey reveals that approximately 0.5% of children born to HIV-negative mothers are HIV-positive, with significantly higher prevalence observed in urban areas, wealthier households, and Copperbelt Province, underscoring critical gaps in current prevention strategies that extend beyond vertical transmission.

The Gist

Imagine Zambia is a massive, bustling garden. For years, the gardeners (health officials) have been incredibly successful at stopping a specific type of weed (HIV) from spreading from the mother plants to the baby plants. They've built strong fences and used excellent tools to ensure that if a mother plant is healthy, her baby should be too. This is called "Prevention of Mother-to-Child Transmission" (PMTCT), and it has worked wonders.

But recently, the gardeners noticed something strange and puzzling. Even though the mother plants tested perfectly healthy, some of the baby plants were still showing signs of the weed.

This paper is like a detective story trying to solve that mystery using a giant snapshot of the entire garden taken in 2024 (the 2024 Zambia Demographic and Health Survey).

The Big Reveal: A Small but Important Mystery

The researchers looked at 13,960 children whose mothers were confirmed to be HIV-negative. They found that 69 of these children (about 0.58%) were actually HIV-positive.

Think of it this way: If you have a classroom of 200 kids, and their moms are all healthy, you'd expect zero kids to have the virus. But here, in a huge school of 14,000, about 7 kids had the virus despite their moms being clean. This proves that the virus is finding a way in through a different door, not the one the gardeners were watching so closely.

Where is the "Weed" Growing? (The Patterns)

The detectives found three very interesting patterns about where these "mystery cases" were happening:

1. The Copperbelt Hotspot:
   Imagine the garden has different sections. One section, called the Copperbelt Province, was a "weed factory." The infection rate there was five times higher than the national average. It's like finding that while the whole garden has a few weeds, one specific patch is overrun. This area is also very industrial (mining) and crowded, suggesting that the virus might be spreading through busy, high-traffic areas.

2. City vs. Country:
   The virus was much more common in city kids than in country kids. City children were nearly 4 times more likely to have the virus than rural children.

   • Analogy: Think of the city as a busy train station where people (and viruses) move fast and mix constantly. The countryside is more like a quiet village where everyone knows everyone, and the virus has a harder time jumping around.

3. The "Rich Kid" Paradox:
   This was the most surprising twist. Usually, we think poor people are at higher risk for diseases because they might lack clean water or medicine. But here, children from wealthier families were actually more likely to be infected than those from the poorest families.

   • Analogy: It's like finding that the people with the fancy, expensive cars are getting more flat tires than the people with old bicycles. Why? Probably because the rich people are driving more, going to more places (like the busy city or the Copperbelt), and interacting with more people. Their wealth put them in the "high-risk zones" more often.

How Did the Virus Get In? (The Suspects)

Since the moms were healthy, the virus didn't come from them. The paper suggests two main suspects:

• The "Dirty Syringe" Suspect (Healthcare Transmission): Sometimes, medical tools aren't cleaned perfectly, or blood products aren't screened well enough. If a child gets a shot or a blood transfusion with contaminated equipment, they could catch the virus. This is like getting a clean plant but accidentally watering it with dirty water.
• The "Bad Neighbor" Suspect (Sexual Abuse): For older children (especially teens), the virus might be coming from sexual abuse or early sexual activity. This is a very sad but real possibility that the study highlights.

What Should We Do? (The Gardeners' New Plan)

The paper argues that the gardeners need to change their strategy. They can't just watch the "mother-to-baby" fence anymore; they need to look at the whole garden.

• Focus on the Hotspots: They need to send extra help and resources to the Copperbelt Province and big cities.
• Check the Tools: They need to make sure every hospital and clinic is using perfectly clean needles and safe blood.
• Protect the Kids: They need to strengthen systems to protect children from abuse.
• Look Deeper: They need to stop assuming that only poor people get sick. The virus is sneaking in through the "rich" and "city" doors too.

The Bottom Line

This study is a wake-up call. It tells us that while we've done a great job stopping the virus from passing from mom to baby, we have a new, hidden problem. The virus is finding other ways to infect children, especially in crowded cities and wealthier areas. To truly win the battle, we need to close all the doors, not just the main one.

Technical Summary

1. Problem Statement

Despite significant progress in Zambia's Prevention of Mother-to-Child Transmission (PMTCT) programs, a persistent public health challenge remains: children testing HIV-positive despite being born to mothers who tested HIV-negative.

• The Gap: Current HIV surveillance and intervention strategies in Zambia are heavily focused on vertical transmission (MTCT). However, the existence of HIV-positive children with HIV-negative mothers suggests undetected non-vertical transmission pathways.
• Potential Mechanisms: These pathways may include healthcare-associated transmission (unsafe injections, contaminated blood products), sexual abuse, or early sexual debut.
• Context: With recent reductions in international funding (e.g., PEPFAR), there is an urgent need for evidence-based targeting of prevention resources beyond the traditional PMTCT cascade.

2. Methodology

• Study Design: Cross-sectional secondary analysis of the 2024 Zambia Demographic and Health Survey (ZDHS 2024).
• Data Source: Publicly available, de-identified datasets (HIV Test Recode and Person Recode) from the DHS Program, collected between January 17 and July 7, 2024.
• Study Population:
  • Inclusion Criteria: Children aged 2–14 years with valid HIV test results linked to a co-resident mother with a valid HIV-negative test result.
  • Exclusion Criteria: Children with inconclusive HIV results, mothers with inconclusive/positive results, or discrepant age data.
  • Final Sample: 13,960 children born to HIV-negative mothers.
• Variables:
  • Outcome: Child HIV status (Positive/Negative).
  • Predictors: Province, residence type (urban/rural), child sex, child age group (2–4, 5–9, 10–14), household wealth quintile, and maternal education level.
• Statistical Analysis:
  • Used survey-weighted estimates to account for the complex two-stage stratified cluster sampling design.
  • Employed design-adjusted Rao-Scott chi-square tests for bivariate associations.
  • Calculated crude Odds Ratios (OR) with 95% Confidence Intervals (CI).
  • Software: Python (v3.14) using pandas, NumPy, statsmodels, and linearmodels.

3. Key Contributions

• First National Assessment: This is the first study to utilize the 2024 ZDHS to quantify the burden of HIV specifically among children born to HIV-negative mothers in Zambia.
• Identification of Non-Vertical Pathways: By isolating this specific subpopulation, the study provides empirical evidence that a significant portion of pediatric HIV cases in Zambia cannot be attributed to maternal transmission, necessitating a shift in prevention paradigms.
• Geographic and Socioeconomic Insights: The study identifies specific high-burden regions (Copperbelt) and counterintuitive socioeconomic patterns (wealth gradient) that challenge standard assumptions about pediatric HIV risk factors.

4. Key Results

• Overall Prevalence: Among 13,960 children of HIV-negative mothers, 69 (0.49% unweighted; 0.58% weighted) tested HIV-positive.
• Geographic Variation:
  • Prevalence varied significantly by province ($p < 0.001$).
  • Copperbelt Province had the highest burden at 2.97% (approx. 5 times the national average).
• Residence:
  • Urban children had significantly higher prevalence (1.10%) compared to rural children (0.26%).
  • Urban residence was associated with a 3.68 times higher odds of HIV positivity ($95\% CI: 2.25–6.03$).
• Wealth Gradient (Paradoxical Finding):
  • A positive association was observed between household wealth and HIV prevalence ($p < 0.001$).
  • Children in the richer quintile had nearly 5 times the odds of being HIV-positive compared to the poorest quintile ($OR = 4.93$).
• Other Factors:
  • Maternal Education: Marginally significant association ($p = 0.04$), with higher prevalence among children of mothers with higher education levels.
  • Sex and Age: No statistically significant associations were found for child sex ($p = 0.10$) or age group ($p = 0.29$), though point estimates suggested slightly higher risks for males and older children (10–14 years).

5. Significance and Implications

• Programmatic Shift: The findings indicate that the current PMTCT-focused paradigm is insufficient. Prevention strategies must expand to address non-vertical transmission risks.
• Targeted Interventions:
  • Copperbelt Province: Requires province-specific responses, including strengthened infection prevention and control (IPC) in healthcare facilities and targeted case-finding in mining/urban communities.
  • Urban Settings: Urban health systems need enhanced surveillance for pediatric HIV.
  • Child Protection: The data underscores the need to strengthen child protection systems to address sexual abuse as a transmission vector.
• Healthcare Safety: The high prevalence in urban and wealthier settings suggests potential issues with healthcare-associated transmission (e.g., unsafe injections or blood screening), warranting audits of medical practices.
• Policy Urgency: These insights are critical for optimizing resource allocation, especially as international funding for HIV programs in Zambia declines.

Limitations Noted: The study is limited by a small number of positive cases ($n=69$), which restricts statistical power and prevents multivariable regression. Additionally, the cross-sectional design precludes establishing causality, and the DHS data lacks specific variables on healthcare exposure or sexual abuse history.