Complete blood cell count as a surrogate CD4 cell marker for HIV monitoring in resource-limited settings

BACKGROUND

A total lymphocyte count (TLC) of 1200 cells/mL has been used as a surrogate for a CD4 count of 200 cells/microL in resource-limited settings with varying results. We developed a more effective method based on a decision tree algorithm to classify subjects.

METHODS

A decision tree was used to develop models with the variables TLC, hemoglobin, platelet count, gender, body mass index, and antiretroviral treatment status of subjects from the University of Alabama at Birmingham (UAB) observational database. Models were validated on data from the Birmingham Veterans Affairs Medical Center (BVAMC) and Zambia, with primary decision trees also generated from these data.

RESULTS

A total of 1189 patients from the UAB observational database were included. The UAB decision tree classified a CD4 count < or =200 cells/microL as better than a TLC cut-point of 1200 cells/mL, based on the area under the curve of the receiver-operator characteristic curve (P < 0.0001). When applied to data from the BVAMC and Zambia, the UAB-based decision tree performed better than the TLC cut-point of 1200 cells/mL (BVAMC: P < 0.0001; Zambia: P = 0.0009) but worse than a decision tree based on local data (BVAMC: P < or = 0.0001; Zambia: P < or = 0.0001).

CONCLUSION

A decision tree algorithm based on local data identifies low CD4 cell counts better than one developed from a different population or a TLC cut-point of 1200 cells/mL.

Early clinical and immune response to NNRTI-based antiretroviral therapy among women with prior exposure to single-dose nevirapine

OBJECTIVE

To determine whether prior exposure to single-dose nevirapine (NVP) for prevention of mother-to-child HIV transmission (PMTCT) is associated with attenuated CD4 cell response, death, or clinical treatment failure in women starting antiretroviral therapy (ART) containing non-nucleoside reverse transcriptase inhibitors (NNRTI).

METHODS

Open cohort evaluation of outcomes for women in program sites across Zambia. HIV treatment was provided according to Zambian/World Health Organization guidelines.

RESULTS

Peripartum NVP exposure status was known for 6740 women initiating NNRTI-containing ART, of whom 751 (11%) reported prior use of NVP for PMTCT. There was no significant difference in mean CD4 cell change between those exposed or unexposed to NVP at 6 (+202 versus +182 cells/microl; P = 0.20) or 12 (+201 versus +211 cells/microl; P = 0.60) months. Multivariable analyses showed no significant differences in mortality [adjusted hazard ratio (HR), 1.2; 95% confidence interval (CI), 0.8-1.8] or clinical treatment failure (adjusted HR, 1.1; 95% CI, 0.8-1.5). Comparison of recent NVP exposure with remote exposure suggested a less favorable CD4 cell response at 6 (+150 versus +219 cells/microl; P = 0.06) and 12 (+149 versus +215 cells/microl; P = 0.39) months. Women with recent NVP exposure also had a trend towards elevated risk for clinical treatment failure (adjusted HR, 1.6; 95% CI, 0.9-2.7).

CONCLUSION

Exposure to maternal single-dose NVP was not associated with substantially different short-term treatment outcomes. However, evidence was suggestive that exposure within 6 months of ART initiation may be a risk factor for poor treatment outcomes, highlighting the importance of ART screening and initiation early in pregnancy.

Use of traditional medicine among pregnant women in Lusaka, Zambia

OBJECTIVE

We studied the prevalence of and predictors for traditional medicine use among pregnant women seeking care in the Lusaka, Zambia public health system.

SUBJECTS

We surveyed 1128 pregnant women enrolled in a clinical trial of perinatal human immunodeficiency virus (HIV) prevention strategies at two district delivery centers.

OUTCOME MEASURES

Postpartum questionnaires were administered to determine demographic characteristics, behavioral characteristics, HIV knowledge, and prior use of traditional medicines.

RESULTS

Of the 1128 women enrolled, 335 (30%) reported visiting a traditional healer in the past; 237 (21%) reported using a traditional healer during the current pregnancy. Overall, 54% believed that admitting to a visit to a traditional healer would result in worse medical care. When women who had used traditional medicines were compared to those who had not, no demographic differences were noted. However, women who reported use of traditional medicine were more likely to drink alcohol during pregnancy, have >or=2 sex partners, engage in "dry sex," initiate sex with their partner, report a previously treated sexually transmitted disease, and use contraception (all p < 0.01). HIV-infected women who reported using traditional healers were also less likely to adhere to a proven medical regimen to reduce HIV transmission to their infant (25% versus 50%, p = 0.048).

CONCLUSIONS

Use of traditional medicine during pregnancy is common, stigmatized, and may be associated with nonadherence to antiretroviral regimens. Health care providers must open lines of communication with traditional healers and with pregnant women themselves to maximize program success.

Prevalence and distribution of HPV genotypes among HIV-infected women in Zambia

We screened 145 HIV-infected non-pregnant women at a tertiary care centre in Lusaka, Zambia. Liquid-based cytology and human papillomavirus (HPV) genotyping with PGMY09/11 biotinylated primers (Roche Linear Array® HPV genotyping test) maximised sensitivity of cytology and HPV assessments. Among high-risk (HR) types, HPV 52 (37.2%), 58 (24.1%) and 53 (20.7%) were more common overall than HPV 16 (17.2%) and 18 (13.1%) in women with high-grade squamous intraepithelial lesions or squamous cell carcinoma (SCC) on cytology. High-risk HPV types were more likely to be present in women with CD4⁺ cell counts <200 μ l⁻¹ (odds ratios (OR): 4.9, 95% confidence intervals (CI): 1.4–16.7, P=0.01) and in women with high-grade or severe cervical cytological abnormalities (OR: 8.0, 95% CI: 1.7–37.4, P=0.008). Human papillomavirus diversity in high-grade lesions and SCC on cytology suggests that HPV 16- and 18-based vaccines may not be adequately polyvalent to induce protective immunity in this population.

Field performance of a thin-layer chromatography assay for detection of nevirapine in umbilical cord blood

PURPOSE

Although cord blood surveillance can measure the effectiveness of nevirapine (NVP)-based programs for the prevention of mother-to-child HIV transmission (PMTCT), it requires the ability to detect nevirapine in plasma. At present, the only validated method is high-performance liquid chromatography (HPLC), a technique poorly suited for most resource-constrained settings.

METHOD

We evaluated the field performance for a simple and inexpensive thin-layer chromatography (TLC) assay for NVP detection. We developed a conditional probability model to compare 2 testing algorithms: HPLC alone, and TLC screening followed by HPLC confirmation of negative results.

RESULTS

When compared to HPLC, sensitivity of TLC was 0.67 (95% confidence interval [CI] 0.49-0.84) and specificity was 0.84 (95% CI 0.69-0.95). In this sample - where overall NVP coverage was 49% - positive predictive value was 0.80 and negative predictive value was 0.72. At baseline with population NVP coverage of 33%, cost per specimen was lower in the TLC-HPLC testing algorithm (40 dollars vs. 50 dollars), and the proportion of false results was acceptable (11%). As population NVP coverage increased, cost-efficiency improved and error rate dropped substantially.

CONCLUSION

TLC is reasonably sensitive and specific for NVP detection. A 2-step testing algorithm incorporating TLC and HPLC provides cost-efficiency at little expense to test performance.

Individual-level predictors for HIV testing among antenatal attendees in Lusaka, Zambia

Despite the availability of antiretroviral prophylaxis, roughly one-fifth of public-sector antenatal patients decline HIV testing in Lusaka, Zambia. We administered a survey to determine individual-level predictors of HIV testing. Of 1064 antenatal attendees approached after pretest counseling, 1060 (>99%) participated. Of these, 686 (65%) agreed to HIV testing. On bivariate analysis controlling for clinic of attendance, women younger than 20 years old (adjusted RR [ARR] = 1.14), unmarried (ARR = 1.14), pregnant for the first time (ARR = 1.14), with lower educational attainment (ARR = 1.15), and with lower income (ARR = 1.14) were all more likely to undergo testing. When HIV risk was considered, women with low self-perceived risk were most likely to undergo HIV testing. As risk perception increased, likelihood for testing decreased (P for trend < 0.001). Although not statistically predictive, we identified prevalent community beliefs that may act as barriers to testing. Because individual-level characteristics were only weakly predictive of HIV testing, future work should concentrate on community-level factors.

Rapid scale-up of antiretroviral therapy at primary care sites in Zambia: feasibility and early outcomes

CONTEXT

The Zambian Ministry of Health has scaled-up human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) care and treatment services at primary care clinics in Lusaka, using predominately nonphysician clinicians.

OBJECTIVE

To report on the feasibility and early outcomes of the program.

DESIGN, SETTING, AND PATIENTS

Open cohort evaluation of antiretroviral-naive adults treated at 18 primary care facilities between April 26, 2004, and November 5, 2005. Data were entered in real time into an electronic patient tracking system.

INTERVENTION

Those meeting criteria for antiretroviral therapy (ART) received drugs according to Zambian national guidelines.

MAIN OUTCOME MEASURES

Survival, regimen failure rates, and CD4 cell response.

RESULTS

We enrolled 21,755 adults into HIV care, and 16,198 (75%) started ART. Among those starting ART, 9864 (61%) were women. Of 15,866 patients with documented World Health Organization (WHO) staging, 11,573 (73%) were stage III or IV, and the mean (SD) entry CD4 cell count among the 15,336 patients with a baseline result was 143/microL (123/microL). Of 1142 patients receiving ART who died, 1120 had a reliable date of death. Of these patients, 792 (71%) died within 90 days of starting therapy (early mortality rate: 26 per 100 patient-years), and 328 (29%) died after 90 days (post-90-day mortality rate: 5.0 per 100 patient-years). In multivariable analysis, mortality was strongly associated with CD4 cell count between 50/microL and 199/microL (adjusted hazard ratio [AHR], 1.4; 95% confidence interval [CI], 1.0-2.0), CD4 cell count less than 50/microL (AHR, 2.2; 95% CI, 1.5-3.1), WHO stage III disease (AHR, 1.8; 95% CI, 1.3-2.4), WHO stage IV disease (AHR, 2.9; 95% CI, 2.0-4.3), low body mass index (<16; AHR,2.4; 95% CI, 1.8-3.2), severe anemia (<8.0 g/dL; AHR, 3.1; 95% CI, 2.3-4.0), and poor adherence to therapy (AHR, 2.9; 95% CI, 2.2-3.9). Of 11,714 patients at risk, 861 failed therapy by clinical criteria (rate, 13 per 100 patient-years). The mean (SD) CD4 cell count increase was 175/microL (174/microL) in 1361 of 1519 patients (90%) receiving treatment long enough to have a 12-month repeat.

CONCLUSION

Massive scale-up of HIV and AIDS treatment services with good clinical outcomes is feasible in primary care settings in sub-Saharan Africa. Most mortality occurs early, suggesting that earlier diagnosis and treatment may improve outcomes.

HIVCorps: using volunteers to rapidly expand HIV health services across Zambia

In 2004, we created HIVCorps, an international volunteer program to involve pre-medical, medical, and public health students in the scale-up of HIV care and prevention services in Zambia. In our first year, we used 27 American and Zambian volunteers to assist with the administrative and logistical aspects of program implementation. Ten volunteers were based in the capital Lusaka; the remaining 17 were stationed across five rural districts. Supervision was provided by local health care providers, district officials, and hospital administrators. In our setting, the use of volunteers has proven feasible and effective for program support. Depending on a program's immediate needs, use of many basic field personnel may be more beneficial than employment of one to two trained clinicians. Formal volunteer programs like HIVCorps should be developed alongside initiatives focused on deploying more specialized, experienced healthcare workers aboard.

Acute and chronic chorioamnionitis and the risk of perinatal human immunodeficiency virus-1 transmission

OBJECTIVE

This study was undertaken to examine the prevalence of acute and chronic chorioamnionitis among women infected with human immunodeficiency virus-1 (HIV-1) and to determine the relative contribution of each to perinatal HIV-1 transmission.

STUDY DESIGN

In 227 HIV-infected women receiving intrapartum/neonatal nevirapine prophylaxis, we examined associations between fetal membrane histology, cord blood interleukin-6 (IL-6), and perinatal HIV-1 transmission.

RESULTS

Acute chorioamnionitis was present in 122 of 227 specimens; chronic chorioamnionitis in 64 of 227. There was a positive correlation between acute chorioamnionitis and labor length (r = 0.208; P = .002), time of ruptured membrane (r = 0.177; P = .008), and cord IL-6 (r = 0.390; P < .001). Chronic chorioamnionitis was associated with high viral load (P = .05) and low cord IL-6 (P < .001). Severe chronic chorioamnionitis was associated with intrauterine HIV-1 transmission (odds ratio [OR] = 7.61; 95% CI = 1.04-85.5), but no correlation was demonstrated between acute chorioamnionitis and vertical transmission.

CONCLUSION

In a setting of high perinatal nevirapine use, acute chorioamnionitis was not associated with vertical HIV-1 transmission. Risk for intrauterine transmission increased significantly when chronic chorioamnionitis was present.

Cost and enrollment implications of targeting different source population for an HIV treatment program

BACKGROUND

Rapid scale-up of antiretroviral therapy (ART) is a worldwide priority, and ambitious targets for numbers on ART have been set. Antenatal clinics (ANCs) and tuberculosis (TB) clinics have been targeted as entry points into HIV care.

METHODS

We developed a conditional probability model to evaluate the effects of ANC and TB clinic populations on ART program enrollment.

RESULTS

To start 1 individual on ART, 3 TB patients have to be screened at a crude program cost of 36 US dollars per patient initiated on therapy. By contrast, 48 ANC patients have to be screened at a cost of US 214 US dollars per patient on therapy. In an incremental analysis in which ANC HIV testing was borne by a program to prevent mother-to-child transmission, recruitment efficiency increased (8 screened per patient starting ART) and cost decreased (114 US dollars per patient on therapy). Absolute numbers starting ART, however, remained fixed. If all 60,000 ANC patients seen yearly in the Lusaka District were screened, 1247 would start ART. Approaching the district's 35,000 annual TB patients would generate 11,947 patients on ART.

CONCLUSION

In areas with high HIV prevalence, targeting chronically ill populations for HIV treatment may have significant short-term benefits in cost savings and recruitment efficiency.

Potential cost-effectiveness of maternal and infant antiretroviral interventions to prevent mother-to-child transmission during breast-feeding

INTRODUCTION

One-third of maternal-to-child HIV transmission occurs during breast-feeding (BF). Several trials are currently evaluating the efficacy of postpartum antiretrovirals to reduce BF transmission.

METHODS

This study used Markov modeling to define the circumstances under which the following interventions would be cost-effective: BF for 6 months with daily infant nevirapine (NVP) prophylaxis; maternal combination antiretroviral therapy (ART) during pregnancy and for 6 months of BF; and maternal combination ART only for women who meet CD4 criteria. Each was compared to: BF for 12 months; BF for 6 months; and formula feeding for 12 months. Strategies were evaluated for a hypothetical cohort of 40,000 pregnant women in sub-Saharan Africa, in the context of available voluntary counseling and testing in antenatal care. Model estimates were derived from the literature and local sources. Sensitivity analyses were performed on uncertain estimates. The perspective used was that of a government health district.

RESULTS

Using base case estimates, BF for 6 months was the economically preferred strategy: it cost 806,995 dollars and generated 446,208 quality-adjusted life-years (QALYs). Providing daily infant NVP cost an additional 93,638 dollars and generated 1183 additional QALYs, but its incremental cost-effectiveness ratio (ICER) of 79 dollars/QALY exceeded the standard willingness to pay (64 dollars/QALY) for most resource-poor settings. Maternal combination ART was potentially very effective but too costly for most resource-poor settings (ICER: 87 dollars/QALY). In order for daily infant NVP during BF to be preferred, it must have >/=44% relative efficacy or cost </=5.00 dollars/mo. If NVP were donated, it would only have to be minimally effective to be the economically preferred strategy. If ART cost </=34.50 dollars/mo, ART to all mothers would become the preferred strategy under our assumption of 82% efficacy.

CONCLUSIONS

Providing antiretrovirals during BF represents a promising alternative, should their effectiveness, and feasibility be proven.

Perceptions toward HIV, HIV screening, and the use of antiretroviral medications: a survey of maternity-based health care providers in Zambia

Mother-to-child transmission of HIV (MTCT) is a major contributor to Zambia's HIV burden. Based on our experience in Zambia, we felt that provider perceptions, knowledge base, and practice patterns toward HIV-positive mothers may pose as significant obstacles to preventing MTCT. Two hundred and twenty-five health care providers throughout Zambia were surveyed in 2002. Providers reported widespread stigma associated with HIV. Physicians (OR = 1.9), providers with research affiliations (OR = 2.3), and those located in Lusaka (OR = 9.0) were more likely to offer HIV testing. Only 30% routinely prescribed antiretroviral treatment (ART) to reduce MTCT. Practitioners from district facilities, those from Lusaka, and those employed at research facilities were more likely to prescribe ART routinely (OR = 2.8, 10.1 and 3.4 respectively). Among those never prescribing ART, most cited a lack of availability (83%). Our results highlight the need for further provider education, critical appraisal of the current system for HIV testing, and widespread distribution of ART.

Rapid testing strategies for HIV-1 serodiagnosis in high-prevalence African settings

BACKGROUND

Nearly 30 million people in sub-Saharan Africa are infected with HIV. While recent commitment of resources from international donors is heartening, rational use of these resources for AIDS prevention and care will require a major scaling up of HIV diagnostic services in affected countries.

OBJECTIVE

This paper considers the various settings and populations in which HIV testing might be implemented, and the goals to be achieved by that testing. It also defines the practical information that public health decision makers in sub-Saharan Africa should have in order to choose an appropriate test or combination of tests for use in a particular clinical setting.

METHODS

Using a conditional probability model and published performance characteristics of rapid HIV serologic tests, the clinical effectiveness and costs of three major testing strategies are evaluated: (1) a single, highly sensitive test; (2) a serial algorithm, where positive results on a first test are confirmed with a more specific second test; and (3) a parallel algorithm, where each specimen is tested with two separate rapid tests, and discordant results are resolved with a third, different rapid test. This analysis was performed in 2003.

RESULTS

We suggest that in any setting, both a serial and a parallel algorithm yield fewer incorrect results than does a single screening assay, but are more costly. A parallel testing algorithm yields fewer incorrect results than does a serial algorithm, but is more costly.

CONCLUSIONS

We suggest that while a parallel testing algorithm has the advantage of avoiding indeterminate results, that strategy may be prohibitively costly for many developing world settings. Furthermore, we suggest that different testing algorithms are appropriate for different clinical settings.

Universal nevirapine upon presentation in labor to prevent mother-to-child HIV transmission in high prevalence settings

OBJECTIVE

To assess the uptake of and adherence to nevirapine to prevent mother-to-child HIV transmission among women of unknown HIV serostatus presenting in labor. We also assessed preliminary efficacy of the approach.

DESIGN

Women of unknown HIV serostatus presenting in labor were offered single-dose nevirapine in a prospective cohort study. Two additional contemporaneous comparison populations were also studied.

METHODS

We measured uptake by counting the number of women that accepted enrollment when offered. We measured adherence with cord blood nevirapine assay. We measured preliminary efficacy with HIV DNA polymerase chain reaction of infant blood spots at 4-6 weeks of life.

RESULTS

Of 1591 women approached in labor, 634 (40%) took up the intervention and received nevirapine, of whom 185 (29%) were HIV infected. Of 179 cord blood specimens from HIV-exposed infants that could be evaluated, 178 (99.4%) had nevirapine detected. This was higher than the 73 of 98 (74%) adherence rate observed in a comparison cohort in which women self-administered nevirapine before presenting to the labor ward (P < 0.001). Of 145 available infant specimens, 17 (11.7%) showed evidence of infection at 4-6 weeks, compared with 12 of 60 (20%) infants born immediately prior to study commencement whose HIV-infected mothers did not receive nevirapine (P < 0.05).

CONCLUSIONS

Nevirapine without HIV testing upon presentation in labor was accepted by two-fifths of women. Because therapy is directly observed, adherence is nearly perfect. Labor ward dosing to enhance nevirapine coverage should be considered as an adjunct to antenatal nevirapine administration for prevention of mother-to-child transmission of HIV.

Detection of nevirapine in plasma using thin-layer chromatography

BACKGROUND

Nevirapine (NVP) is widely prescribed in resource-poor settings to pregnant women for treatment and prevention of HIV infection. High rates of misreported adherence, however, have compelled clinicians to find alternative methods to ensure systemic drug exposure. This report describes a fast, inexpensive thin-layer chromatography (TLC) method to detect the presence of NVP in human plasma.

METHODS

Human plasma was spiked with various concentrations of NVP. NVP was subsequently isolated using solid-phase extraction and visualized with TLC. Clinical samples with NVP concentrations predetermined by high-performance liquid chromatography were used to validate the TLC method.

RESULTS

NVP was detected at concentrations as low as 60 ng/mL. The lower limit of detection was set at 100 ng/mL due to the clear spot definition at this concentration. The turnaround time for assay results averages several hours, and costs associated with the assay are considerably below standard drug quantitation techniques.

CONCLUSION

TLC provides a rapid, sensitive, and economical tool to qualitatively measure NVP in plasma. This method offers clinicians in resource-poor settings an alternative approach for measuring adherence, particularly in developing-world regions where NVP use is common and there is an immediate need to prevent mother-to-child HIV transmission.

Personal risk perception, HIV knowledge and risk avoidance behavior, and their relationships to actual HIV serostatus in an urban African obstetric population

One quarter of pregnant women in Zambia are infected with HIV. Understanding how knowledge of HIV relates to personal risk perception and avoidance of risky behaviors is critical to devising effective HIV prevention strategies. In conjunction with a large clinical trial in Lusaka, Zambia, we surveyed postpartum women who had been tested for HIV but did not know their status before undergoing the questionnaire. Of 858 women for whom complete data were available, 248 (29%) were HIV infected. Women 22 years of age or older (adjusted odds ratio [AOR], 1.7; 95% confidence interval [CI], 1.1-2.5), women reporting > or =2 sexual partners in their lifetime (AOR, 1.8; 95% CI, 1.3-2.5), and women reporting a history of a sexually transmitted infection (AOR, 2.7; 95% CI, 1.7-4.3) were more likely to be HIV infected. Having had > or =2 lifetime sexual partners was a marker for perception of high personnel risk for HIV infection (AOR, 1.5; 95% CI, 1.1-2.1). However, there was no relationship between perceived risk of HIV infection and actual HIV status. In fact, 127 (52%) of 245 women who stated that they were at no or low risk for HIV infection were HIV infected. Living in an area of high HIV seroprevalence like Zambia seems to be the greatest risk factor for infection in unselected pregnant women. Before significant inroads can be made in decreasing the incidence of HIV infection among pregnant women, population-based strategies that involve men must be implemented.

Timing of the maternal drug dose and risk of perinatal HIV transmission in the setting of intrapartum and neonatal single-dose nevirapine

CONTEXT

Single-dose intrapartum and neonatal nevirapine (NVP) reduces perinatal HIV transmission and is in increasingly common use throughout the developing world.

OBJECTIVE

We studied risk factors for perinatal transmission in the setting of NVP.

DESIGN AND SETTING

A prospective cohort study at two public obstetrical clinics in Lusaka, Zambia.

PATIENTS AND METHODS

In a volunteer sample of HIV-infected pregnant women and their newborns, the women received a 200 mg oral dose of NVP at the onset of labor; their infants received 2 mg/kg of NVP syrup within 24 h of birth. The main outcome measure was the infant HIV infection status at 6 weeks of life, determined by DNA polymerase chain reaction.

RESULTS

Only 31 of 278 (11.2%) infants were infected at 6 weeks. In logistic regression, viral load exceeding the median [adjusted odds ratio (AOR), 3.1; 95% confidence interval (CI), 1.1-8.7] and 1 h or less elapsing between NVP ingestion and delivery (AOR, 5.0; 95% CI, 1.8-14) were associated with transmission. Women delivering within 1 h of NVP ingestion had a lower mean drug concentration (351 versus 942 ng/ml; P<0.001) and were more likely to have a 'sub-therapeutic' NVP level of less than 100 ng/ml (56 versus 20%; P<0.001) than those who delivered more than 1 h post-ingestion. However, concentrations <100 ng/ml were not more likely to be associated with transmission than concentrations > or = 100 ng/ml (12.9 versus 11.7%; P=0.8). We did not identify a threshold concentration below which risk of transmission increased.

CONCLUSIONS

We confirmed low perinatal transmission rates with single-dose NVP. At least 1 h of pre-delivery NVP prophylaxis was a critical threshold for efficacy.

Mother-to-child HIV transmission prevention in Thailand: physician zidovudine use and willingness to provide care

We conducted a mail survey of Thai physicians involved in obstetric care to assess attitudes and practices regarding zidovudine use during pregnancy and willingness to provide care for HIV-infected women in 1999. Of 845 respondents, 57% reported using perinatal zidovudine prophylaxis, an increase from 20% reported in 1997. Highest failure-to-use rates (52%) were among the respondents from Central and Southern Thailand and lowest failure rate was among those from the North (37%). Predictors of failure to use zidovudine in a multivariable logistic regression analysis were not knowing a source from which to obtain zidovudine (odds ratio [OR]=3.1), working in smaller hospitals (district/provincial/private hospitals) (OR=2.0), being from Eastern/Central/Southern Thailand (OR=1.4), unwillingness to perform caesarean section delivery on a HIV-positive women (OR=1.8), having provided antenatal care to fewer than 100 women in 1998 (OR=1.7), and unfamiliarity with Pediatric AIDS Clinical Trial Group 076 protocol (OR=2.9). A number of respondents described themselves as unwilling to perform pelvic examinations (15%), vaginal delivery (29%), or caesarean sections (37%) on HIV-infected pregnant women. About 39% of the respondents advocated elective terminations of pregnancy for HIV-infected women. Our survey indicates an increasing willingness of Thai physicians to use antiretroviral therapy to prevent mother-to-child HIV transmission and to provide obstetric care to HIV-infected women. However, availability and affordability remained major barriers to more widespread antiretroviral use in 1999.