Background Antenatal clinic (ANC) HIV sentinel surveillance has provided key data to inform health policy, especially between population-based surveys. HIV prevalence among teenage attendees has been used as a proxy for incidence in Zambia, where incidence data are lacking. We present trends in HIV prevalence among ANC attendees in Zambia since 1994.

Methods We assessed HIV prevalence in 15–39 year-old women accessing ANC in convenience sample-based cross-sectional surveys in 21 sites in 1994, 1998, 2002, 2004, 2006, 2008 and 2011. Anonymous blood was tested from consecutively-enrolled women on their first ANC visit; age, residence, parity and gravidity were also noted. We evaluated trends in HIV prevalence.

Results Among 8,222 women surveyed in 2011, HIV prevalence remained essentially unchanged from 2008 (16.2% vs. 16.4%, p = 0.74), but declined from a peak of 19.6% in1994 (p < 0.001). Similarly, HIV prevalence stagnated among 1,661 15–19 year-olds between 2011 and 2008 (7.9% vs. 8.3% p = 0.66) but declined from its peak of 13.9% in 1994 (p < 0.001). Declines in HIV prevalence among rural residents have been modest (11.8% peak in 2004 vs. 10.4% in 2011 [n = 2904], p = 0.09). HIV prevalence among 5,318 urban women declined from 22.8% in 2008 to 19.6% in 2011 (p < 0.001), the lowest level since a peak of 27.6% in 1994 (p < 0.001). Among 904 15–19 year-old urban women HIV prevalence declined in 2011 (10.1%) from a peak of 16.4% in 2002 (p < 0.001), but was essentially unchanged compared to 2008 (11.1%, p = 0.42).

Conclusions HIV prevalence declined in urban ANC attendees from 2008–2011 and in all surveyed ANC attendees compared to 1994. Results suggest success in prevention activities over the past 17 years, especially among urban women. Declines in prevalence have slowed recently, likely reflecting increased treatment-related survival possibly combined with decreases in incidence. Incidence and survival data are needed to fully understand these data.