Item #1: Pregnancy and False Positive HIV Tests

[Hafford] did not know that pregnancy itself can cause a false positive HIV test

Farber states that pregnancy itself can cause a false positive result. She supplies no supporting reference.

A properly conducted HIV-test protocol (which involves at least two HIV tests) has a very small chance of giving a false positive, irrespective of pregnancy status. Farber alleges that Hafford's HIV-test was carried out incorrectly. If this was the case, medical negligence is a different matter to whether HIV tests carried out according to protocol are accurate in pregnant women. HIV tests were highly accurate from the time they were developed in 19842 and have become much more accurate over time as the underlying technology has evolved. HIV tests are amongst the most accurate available in medical science. For more on testing see Mirken (2001)3. For a more technical discussion see Coon (2000)4 . Incidentally, the testing protocol of the PACTG 1022 trial, to which Farber refers, required multiple HIV tests and regular viral load counts. Farber states that Hafford was only tested once. Assuming Farber is right, then Hafford's doctor did not follow the protocol. We, however, are not privy to Hafford's medical records and therefore cannot know if Farber's allegation of Hafford having only one test is correct. Was Harper's privy to this? Consequently, was the allegation fact-checked?

The risk of a false positive HIV test is well documented in the scientific literature:

• “False-positive ELISA [antibody] test results can be caused by alloantibodies resulting from transfusions, transplantation, or pregnancy, autoimmune disorders, malignancies, alcoholic liver disease, or for reasons that are unclear...As the number of women being screened has increased, the proportion of false-positive and ambiguous (indeterminate) test results has increased and the positive predictive value (PPV) [a mathematical expression of the probability that a positive test is a true positive] of the standard HIV test has decreased” [1]
• “potential causes of indeterminate Western blots for HIV-1...included...HLA antibody cross-reactivity, and current pregnancy...All but one of the cases who demonstrated anti-class I HLA reactivity were multiparous women [pregnant more than once]” [2]
• A report in an American Medical Association newsletter: “New York health officials are facing an unexpected hurdle in their efforts to identify newborns whose mothers may have exposed them to HIV in utero. A state regulation went into effect in August 1999 that requires hospitals and birthing centers to test within 48 hours all women who present for labor and whose HIV status is unknown. Under the new requirements, babies are routinely screened at birth with the same test if their mothers decline it.

  But such tests are not always accurate and used appropriately. Preliminary health department data for August through November 1999 showed that screening with a rapid HIV test identified 32 HIV-positive women. After using a confirmatory test, however, 17 women were later determined to be HIV-negative. In the meantime, 13 of the 17 babies were started on preventive antiretroviral treatment. Babies treated with zidovudine, the standard preventive therapy, can develop anemia. The long-term effects of the drug are unknown…A pilot study conducted for research purposes last summer [1999] at a Bronx hospital found an "unacceptably high number" of false-positives, said Montefiore Medical Center epidemiologist Mayris P. Webber, DrPH, who was involved in the study. "One should be cautious in interpreting the results of a small pilot study of 100 women," she said, "but we did find, nonetheless, that out of three positive tests, two were false-positives."

  The predictive value of the test in a low prevalence population can be poor, she said…A study by the New York City Dept. of Health found a 45% false-positive rate for SUDS, said Sara T. Beatrice, PhD, the department's director of retrovirology and immunobiology” [3]

• A test kit manufacturer writes: “nonspecific reactions may occasionally be seen in specimens from people who have prior pregnancy, blood transfusion, or exposure to human cells or media containing cultured HIV antigen…when the ELISA is used to screen populations with a low prevalence of HIV infections, nonspecific reactions may be more common than specific reactions” [4]
• “The medical records of all rapid HIV-ELISA-positive gravidas [pregnant women] that delivered at our hospital between January 2000 and October 2001 were retrieved...The results of the Western blot tests were also retrieved and correlated to the ELISA results, across varying maternal characteristics....A total of 69 patients had a positive rapid HIV-ELISA out of 9,781 deliveries. Of those, 26 were confirmed as HIV infected by Western blot (overall HIV prevalence: 0.27%, ELISA-positive predictive value: 37.7%). The subgroup prevalence of HIV and positive predictive value of ELISA were 1.53 and 75% among Caucasians; 2.43 and 82.6% among African-Americans; and 0.05 and 9.8% among Hispanics, respectively…The positive predictive value of rapid HIV-ELISA during pregnancy varies widely, depending on maternal race/ethnicity and sexual behavior. The routine disclosure of rapid intrapartum HIV serum screening results prior to Western blot confirmation should be avoided in very low-risk populations. [and since there is generally no further confirmation of a positive Western Blot, it is actually impossible to tell whether these are true positives or not]” [5]
• “False-positive HIV ELISAs occur because of human or technical errors associated with doing the tests or because of antibodies that coincidentally cross-react with HIV or nonviral components in the tests…Notable causes of false-positive reactions have been anti-HLA-DR antibodies that sometimes occur in multiparous women and in multiply transfused patients.” [6]
• “Most patients (68 to 89%) from low risk groups (prevalence of 0.1% or less) who show reactivity on screening tests will have false-positive results…The predictive value of a positive ELISA varies from 2% to 99%…One notable association with false positive ELISA reactivity in some commercial preparations has been patients with anti-HLA-DR4 antibodies, most often multiparous [having experienced one or more births] or multiply transfused patients...the Western blot method [used to confirm these false positive ELISA tests] lacks standardization, is cumbersome, and is subjective in interpretation of banding patterns.” [7]
• “We selected the 20 most strongly [indeterminate or atypical Western Blot] reactive samples for further evaluation…Atypical WB [Western Blot] patterns in 19 of 20 of our donors remained substantially the same over time…our data show that the presence of p24 alone in WB should not be regarded as a false positive without subsequent testing of the individual…All study donors had normal immune status…[2] donors were multiparous females [multiple children], and one other probably had received a blood transfusion…we observed a large proportion of individuals who had either lived or worked on dairy farms (6/16) and frequently drank unpasteurized cows’ milk (7/16)…undefined autoimmune phenomena [such as multiple pregnancies], bovine exposure, or cross-reactivity with other human retroviruses could be possible causes for consistently reactive HIV immunologic assays” [8]

1. Doran TI et al. False-Positive and Indeterminate Human Immunodeficiency Virus Test Results in Pregnant Women. Arch Fam Med. 2000 Sep/Oct; 9: 924-9.
2. Celum CL et al. Risk factors for repeatedly reactive HIV-1 EIA and indeterminate Western blots: a population-based case control study. Arch Intern Med. 1994 May 23; 154: 1129-37.
3. Shelton DL. New York state physicians slam rapid perinatal HIV test. amednews.com. 2000 Mar 13. ama-assn.org/amednews/2000/03/13/hlsa0313.htm.
4. Vironostika HIV-1 plus O microelisa system. Biomérieux. 2003 Jun 5.
5. Zacharias NM et al. High false-positive rate of HIV rapid serum screening in a predominantly hispanic prenatal population. J Perinatol. 2004 Aug 19.
6. Proffitt MR et al. Laboratory diagnosis of human immunodeficiency virus infection. Inf Dis Clin North Am. 1993; 7: 203-19.
7. Steckelberg JM et al. Serologic testing for human immunodeficiency virus antibodies. Mayo Clin Proc. 1988; 63: 373-9.
8. Dock NL et al. Evaluation of atypical human immunodeficiency immunoblot reactivity in blood donors. Transfusion. 1988; 28: 412.