The study examined all HIV testing done in labor or postdelivery in the 2 hospitals in 2000. Results from another study suggest that approximately 7% of women giving birth at these 2 hospitals had not received HIV testing during pregnancy and were thus eligible for testing at the time of delivery. 13 The protocols developed for testing at delivery were different in the 2 hospitals. Hospital A instituted running batches of standard HIV ELISAs 7 times a week (6:30 am each day Monday through Friday, late afternoon Friday, and Saturday noon). Results were reported by telefax if they were negative, and if positive, the responsible physician was to be paged. In July 2000, the protocol in hospital A was changed to doing SUDS tests on request. Hospital B initiated SUDS testing from the start. All SUDS tests were processed in the hospitals' central laboratories and were run as individual tests. The protocol in each hospital stipulated that the laboratory was to be alerted when a specimen was being sent so that the SUDS test kit could be removed from the refrigerator to allow for warming to room temperature. Serum samples were used in hospital B, but in hospital A, plasma samples were requested to eliminate the time required for blood to clot prior to testing. Both hospitals had technicians on call for processing of specimens received during evening hours and during the daytime on weekends, but neither hospital was able to have complete 24-hour coverage. During times when laboratory services were not available, specimens were processed when the laboratory next opened. In hospital A, the specimen was usually added to the batch of ELISAs run at 6:30 am, whereas in hospital B, a SUDS test was used.
Approval for the study was obtained from the institutional review boards of the 2 hospitals. Eligible subjects were women who were in labor at the time of hospital admission and who had HIV testing performed. Counseling was provided and consent obtained before testing was done. Demographic data and the date and time of each event (hospital admission, rupture of membranes, and birth) were abstracted from medical records, and laboratory records were reviewed to obtain the time of the arrival and reporting of results of each specimen.
Of the 100 tests done at hospital A, 64 were ELISAs and 36 were SUDS. In hospital B, all 56 tests were SUDS. Seventy-five percent of the women had not received HIV testing earlier in pregnancy, 9% had not attended prenatal care, an additional 8% had fewer than 4 prenatal visits, 10% had a history of illicit drug use, and 4% had acquired a sexually transmitted infection during pregnancy. One SUDS test result was positive, but the woman tested had received a false-positive HIV test in a previous pregnancy; thus, no interventions were done, and the result was quickly confirmed to be a false-positive.
Comparison of Protocols Using 2 Different Laboratory Tests at Hospital A
The time taken before the results were available for the 2 different tests used in hospital A is summarized in Table 1. After the change from running a daily batch of ELISAs to processing SUDS tests on request, there were significant decreases in both the time taken for samples to get to the laboratory and the time taken in the laboratory. The total time taken from hospital admission to the reporting of results dropped from a mean of 35.3 hours to 8.0 hours (P < 0.0001). There was great variability in the amount of time taken when ELISAs were being used, but the variability in intervals was much less with the SUDS protocol.
The implications for decreasing the time intervals are evident in Figure 1. After the SUDS was introduced in hospital A, significantly more women had results available prior to rupture of membranes (36%) and delivery (64%) compared with when the ELISA protocol was being used (2% and 8%, respectively), and when using the SUDS, all except 1 woman had results reported within 12 hours of giving birth compared with only 22% when the ELISA was used (P < 0.0001). For each of the protocols, nearly all results were available within 48 hours of delivery as stipulated in the state law.
Comparison of Protocols Using the Single Use Diagnostic System at the 2 Hospitals
Comparing the use of the SUDS in the 2 hospitals, the mean intervals of time taken between hospital admission and samples arriving at the laboratories were similar (see Table 1); however, the interval between arrival at the laboratory and the reporting of results was much shorter in hospital A than hospital B. Thus, the total time before results were available was shorter for hospital A than for hospital B (a mean of 8.0 hours vs. 16.3 hours, P < 0.05). The more rapid turnaround time in hospital A meant that results were more often available prior to critical events (see Fig. 1). For example, 64% of results were available prior to delivery in hospital A compared with only 38% in hospital B (P < 0.05).
Specimens obtained in the middle of the night were handled differently in the 2 hospitals: in hospital A, if a sample had waited overnight, an ELISA was done and was therefore excluded from the analysis, whereas in hospital B, a SUDS was used. To determine if this difference might explain differences between the hospitals, we repeated the analyses excluding the subjects who were admitted outside of protocol hours (ie, night and weekend hours specified for the 2 hospitals). This, however, produced little change in the results: for the 36 women who were admitted during protocol hours in hospital A, the total time between admission and reporting of results was 7.9 hours compared with 14.9 hours for the 40 women who were admitted during protocol hours in hospital B (P = 0.06). As before, this difference was largely a result of differences in the times taken in the laboratories.
In hospital B, 17 of the 56 blood specimens were in the laboratory for more than 12 hours before the results were reported. For 10 of these, the women had been admitted on the weekend, 3 others were admitted after 9:00 pm on a weekday, and 2 were admitted after 4:00 pm. The remaining 2 specimens were those that had taken the longest time. For both of these, delivery had occurred within 2 hours of admission and blood had been drawn from the infants, but there were substantial delays in obtaining parental consent to process the tests.
These results demonstrate that protocols for HIV testing of women in labor can successfully ensure that results are available within 48 hours of delivery, which, in this instance, satisfied the requirements of the Connecticut state law. Nevertheless, the study emphasizes that if results are to be available early enough to have the greatest potential for preventing transmission of HIV infection, a rapid test needs to be used. Once the SUDS test was used in hospital A, more than one-third of the results were reported before rupture of membranes and almost two-third of the results were back before delivery. In these instances, if a result had been positive, antiretroviral prophylaxis could possibly have been started before delivery and obstetric measures might have been taken to decrease the risk of transmission. Perhaps most important, given the possibility that it may be critical to start prophylaxis of an HIV-exposed baby within the first 12 hours of birth, nearly all the SUDS results were available prior to this time, but when the ELISA was used at the same hospital, less than 25% of the results were reported by 12 hours of age.
The results of the study also emphasize how important it is that procedures are implemented for ensuring that specimens get to the hospital laboratory as quickly as possible and are processed and reported by the laboratory in a timely fashion. In this study, we were unable to understand fully why there were delays in providing results; however, the data from hospital B suggest that the protocol for processing specimens in the evenings and on weekends was not fully used. It appears that in this hospital, the emphasis was more on satisfying the state law, which stipulated that results should be reported within 48 hours of birth. In hospital A, there was a greater attention to making sure that results were available as expeditiously as possible; for example, the laboratory technologist was paged to notify him or her that a sample was being sent, the SUDS test kit was removed from the refrigerator to allow for warming to room temperature, and a plasma sample was used to eliminate the time needed for blood to clot. The difficulty of processing specimens at all times does, however, raise the question of whether it might be more appropriate in some hospitals to have point-of-care testing done by trained clinicians in the delivery suite. Such testing is governed by the Clinical Laboratory Improvements Act of 1988 (CLIA), which requires strict attention to proficiency testing of laboratories and training and competency testing of personnel performing the testing to ensure the reliability of the results. The SUDS test has been widely viewed as not appropriate for point-of-care testing because of difficulties in reading the test and the potential for false-positive results. 12 OraQuick, however, is more reliable and considered appropriate for point-of-care testing, and in January 2003, the US Department of Health and Human Services announced expanded access to OraQuick through a CLIA waver, thus opening the way for more widespread use of the test in point-of-care testing. A recent study done in the labor units of Chicago hospitals demonstrated that when labor unit staff performed the tests, the median turnaround time was 45 minutes compared with 3 hours when specimens were sent to a central laboratory. 14 Another study, however, has shown that when OraQuick was performed by nonlaboratory personnel without any prior training, there was an 11% error rate, which emphasizes the need for intensive training and continued monitoring of the use of these tests when performed outside of the laboratory. 15 For some hospitals, it might be more appropriate to process OraQuick in a laboratory. Although strict attention is required for the processing of the test, it is easier to read than the SUDS and therefore does not require the same amount of experience and could be conducted in laboratories that are open continuously, such as chemistry and hematology laboratories. As noted by the results of this study, as long as there is critical attention to the procedures needed to ensure rapid turnaround, results should be available within an appropriate time interval to ensure institution of treatment if it is needed.
Initiating protocols for rapid HIV testing in labor can help to decrease the risk of perinatal transmission for those found to be infected, while at the same time, a negative test result obviates the clinician's impulse to provide prophylaxis empirically to a woman considered to be at high risk of being infected—an approach taken by some clinicians. There are substantial complexities, however, associated with HIV counseling and testing in labor, and it will take collaboration between clinicians, laboratory personnel, and hospital administrators to ensure that testing procedures are conducted in an appropriate, expeditious, and reliable manner. Although this study focused on the use of rapid HIV tests for women in labor, this obviously is not the preferred time to be counseling a woman about HIV and obtaining consent for HIV testing. All efforts should be made to increase the rate of testing earlier in pregnancy.
1. Ioannidis JPA, Abrams EJ, Ammann A, et al. Perinatal transmission of human immunodeficiency virus type 1 by pregnant women with RNA virus loads <1000 copies/ml. J Infect Dis. 2001; 183:539–545.
2. Peters V, Thomas P, Marinovich A, et al. Missed opportunities for perinatal HIV prevention in New York City. Presented at the XIII International AIDS Conference, Durban, South Africa, July 2000.
3. Guay LA, Musoke P, Fleming T, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial. Lancet. 1999; 354:795–802.
4. Saba J. The results of the PETRA intervention trial to prevent perinatal transmission in sub-Saharan Africa. Symposium Session 8. Presented at the Sixth Conference on Retroviruses and Opportunistic Infections, Chicago, January–February 1999.
5. Landesman SH, Kalish LA, Burns DN, et al. Obstetrical factors and the transmission of human immunodeficiency virus type 1 from mother to child. N Engl J Med. 1996; 334:1617–1623.
6. The European Collaborative Study. Vertical transmission of HIV-1: maternal immune status and obstetric factors. AIDS. 1996; 10:1675–1681.
7. The International Perinatal HIV Group. Duration of ruptured membranes and vertical transmission of HIV-1: a meta-analysis from 15 prospective cohort studies. AIDS. 2001; 15:357–368.
8. Wade NA, Birkhead GS, Warren BL, et al. Abbreviated regimens of zidovudine prophylaxis and perinatal transmission of the human immunodeficiency virus. N Engl J Med. 1998; 339:1409–1468.
9. Wade NA, Birkhead GS, French PT. Short courses of zidovudine and perinatal transmission of HIV [letter]. N Engl J Med. 1999; 340:1042–1043.
10. Bulterys M, Orloff S, Abrams E, et al. For the Perinatal Aids Collaborative Transmission Study (PACTS): impact of zidovudine post-perinatal exposure prophylaxis on vertical HIV-1 transmission: a prospective cohort study in 4 U.S. cities. Presented at the Second Conference on Global Strategies for the Prevention of HIV Transmission from Mothers to Infants, Montreal, September 1999.
11. Minkoff H, O'Sullivan MJ. The case for rapid HIV testing during labor. JAMA. 1998; 279:1743–1744.
12. Branson BM, Del Rio C, Larrabee S, et al. Rapid fingerstick testing: a new era in HIV diagnostics. AIDS Clin Care. 2003; 15:19–23.
13. Forsyth BWC. Test now or test later: Connecticut's experience with mandatory HIV testing of newborn infants. Presented at the National HIV Prevention Conference, Atlanta, July 2003
14. Cohen MH. Branson B, Olszewski Y, et al. Effective point-of-care rapid HIV testing at labor and delivery. Presented at the 10th Conference on Retroviruses and Opportunistic Infections, Boston, February 2003.
15. Granade TC, Parekh BS, Phillips SK, et al. Performance rapid HIV antibody detection assays by non-laboratorians. Presented at the 19th Annual Clinical Virology Symposium, Clearwater Beach, FL, April 2003.