A 36-year-old man was admitted to the intensive care unit after being found confused and febrile on the street. Admission evaluation revealed evidence of multiple trauma, waxing and waning mental status, and leukocytosis; a urine toxicology screen was positive for opiates and cocaine.
While placing a central line in this patient, the intensivist stuck himself with the hollow-bore needle being used for the procedure. He experienced slight bleeding but did not recall any visible blood on the device. According to state law where the case occurred, written informed consent is required for HIV testing in the source patient. However, the patient refused to sign the consent form and became angry because he felt the healthcare team was threatening him. During the next 24 hours, he became increasingly obtunded and required intubation for airway control. The only person visiting him in the hospital was his girlfriend; no healthcare proxy or family member was available.
How would you manage the intensivist who sustained the injury? Would you give postexposure prophylaxis (PEP)? If so, what specifically would you give, and for how long? How would you determine whether the source patient is HIV-infected? Would you perform surrogate testing (for example, obtain a CD4-cell count) to assess this? Would you consider testing him without written informed consent? Would you ask his girlfriend to provide consent for him? If he is tested, would you give him the results after he recovers if the test is negative? If the test is positive?
Addendum: When we originally published this case, we asked readers what they would do (see responses below). Now, we have updated the case to provide the opinions of three experts.
According to CDC surveillance data, the average risk for HIV infection following percutaneous exposure to HIV-infected blood is approximately 0.3% (Am J Med 1997; 102:9). An exacerbating factor in this case is the involvement of a hollow-bore needle that (presumably) had been intravascular in the source patient; thus, blood might have been present in the lumen of the needle, even if it was not visible on the needle's exterior. Most guidelines and many experts would recommend PEP for such an exposure, based on a single case-control study suggesting that it leads to a >80% reduction in the likelihood of HIV infection following a needlestick injury (N Engl J Med 1997; 337:1485).
In an ideal situation, the source individual would agree to HIV testing, and the results would determine the need for PEP. For example, if rapid testing for HIV antibodies was available and the results were negative, PEP could be deferred, assuming the source was not at risk for being in the acute “window period” of HIV infection (when viremia levels are extremely high, but antibodies are not yet present). If rapid testing was not available, PEP could be initiated while the source patient was tested conventionally (ELISA with confirmation from a Western blot or an immunofluorescent assay) — and then if a negative result was obtained, the PEP could be discontinued (again, with the same caveats regarding acute infection risk).
Laws around HIV testing and consent are evolving. CDC guidelines published in 2006 recommended dropping the requirement for written consent, to destigmatize and facilitate such testing. Although state laws are increasingly coming into concordance with CDC recommendations, such laws differ widely from state to state (Ann Intern Med 2009; 150:263). Even in states that have adopted “opt-out” HIV testing policies and dropped the requirement for written consent, the situation remains extremely complicated and relatively untested against legal challenge: If the source patient is unwilling to be tested, there are no ethical options to get around the refusal. Surrogate testing (of CD4-cell counts, viral loads, or p24 antigenemia) violates the spirit of informed consent for HIV testing and is not considered acceptable. The source patient's girlfriend cannot provide informed consent unless she has been specifically designated as his healthcare proxy.
The clinician then is left to balance the likelihood of HIV acquisition in this case against the potential toxicities of PEP. In the current HIV treatment era, the toxicity of a 28-day course of PEP is significantly less than it was in the not-too-recent past, and therefore the risk–benefit ratio for PEP is improving. Given the lack of data showing that a three-drug regimen for PEP is superior to a two-drug one, I would provide the intensivist with a 28-day course of tenofovir/FTC (Truvada) to be started as quickly as possible after the exposure, with baseline and follow-up ELISA testing at 4 to 6 weeks, 12 weeks, and 24 weeks.
In addition, I would consider the possibility that the intensivist was exposed to hepatitis B or C virus (HBV and HCV, respectively) from the needlestick. If the intensivist has somehow escaped HBV vaccination, a first dose of vaccine should be given immediately; if the source patient has evidence of circulating HBV surface antigen, hepatitis B immune globulin, too, could be considered for the intensivist. HCV testing of the source should be performed, and baseline and serial EIA testing of the intensivist should be considered to assess for acute HCV seroconversion, which would warrant early treatment with pegylated interferon. Notably, hepatitis testing in the source patient would not be considered a surrogate for HIV testing in the same way that testing for viral load, CD4-cell count, or p24 antigenemia would be.
Any version of involuntary testing in this scenario is inappropriate — medically, legally, and ethically. Individual privacy rights do not operate in isolation from medical and public health needs, but a provider's desire to know a piece of information is very different from the need for, or even usefulness of, that information — in this case, the HIV status of the patient.
The patient did not, and does not, have the legal capacity to consent to an HIV test, and without a medical power of attorney, the girlfriend has no legal authority to consent on his behalf. Equally important, however, is that the testing under consideration would not provide information sufficiently conclusive to alter the recommended course of prophylaxis.
None of the tests suggested could rule out HIV infection or the possibility of exposure. A negative antibody test does not rule out recent infection. A CD4-cell count is useless, because it reveals immunocompetence or suppression, not the cause, and many HIV-positive individuals retain healthy immune systems for at least a short period following infection. Even a negative viral-load test is not conclusive in ruling out the presence of HIV.
The risk for infection from the needlestick accident described is extremely low, even if the exposure involved HIV-infected blood. That said, a 4-week course of tenofovir/FTC (Truvada), with its relatively tolerable side effects, would generally be worthwhile to diminish the remote risk for infection. Although the scenario makes no reference to the HIV status or testing of the intensivist, he himself should receive initial and repeat HIV testing.
This recommendation for PEP should be considered in the context of current HIV therapy trends, in which growing numbers of doctors are calling for lifelong antiretroviral therapy (ART) soon, or even immediately, after an individual is infected with HIV. If physicians and researchers thought a short course of treatment was too toxic or risky to endure without proof of HIV exposure, would they be recommending earlier initiation of ART to their patients in advance of clear clinical need and conclusive evidence that it improves long-term health and survival?
Finally, if the patient were tested, not revealing that fact would be unethical, regardless of the test result. However, the question is irrelevant because involuntary testing itself would be unethical under the circumstances described.
- Lisa Lehmann, MD, PhD, MSc
The ethics of testing patients for HIV without their consent must be distinguished from the legal question of whether the law requires informed consent. From an ethical perspective, even in states with no requirement for written informed consent, we would still almost always need to obtain verbal consent from patients before testing them for HIV.
Consent may be waived if not knowing a patient's HIV status would cause serious harm to another person. In the present case, however, not knowing the patient's HIV status does not pose a threat of serious harm to the intensivist. The intensivist can prevent infection by taking PEP without knowing the patient's HIV status. In doing so, the intensivist may incur unnecessary toxicity and adverse effects from the medications. This potential harm must be balanced against the harm to the patient of testing without consent.
In this case, I see no compelling reason to test the patient without his consent. Testing the patient would not be for diagnostic purposes, to provide him with appropriate medical care, but rather for the intensivist's benefit. Furthermore, testing is not absolutely necessary to protect the health of the physician and has the potential to harm the patient. If the intensivist is concerned about his risk for HIV infection, he can begin PEP immediately, and when the patient regains his capacity to provide consent, he can be asked again about testing. However, even if the patient were tested after the needlestick, he could be in a window period during which he could transmit the virus without testing positive. Thus, a negative HIV test for the patient would not guarantee that the intensivist has no risk for HIV infection. If the patient never regains the capacity to consent because he dies, postmortem testing would be reasonable even without consent because concerns about potential harm to him would have disappeared.
Surrogate testing for HIV is ethically unjustified in that it represents a duplicitous attempt to circumvent both the patient's autonomy and state law. It is also clinically problematic because a CD4-cell count is unreliable in critically ill patients and therefore should not be used to guide clinical decisions.
The patient's girlfriend in this case is not an appropriate surrogate to provide informed consent for testing. Some states that require written informed consent mandate that the consent be from the patient, not from a surrogate. Even if the girlfriend could legally provide consent, disregarding her own motivations for wanting the patient tested might be difficult for her.
The intensivist's preferences with regard to PEP should guide the decision about how to proceed. The risk for HIV transmission from needlestick injuries should be explained, as well as the increased risk associated with a needlestick involving a large-bore needle that was in a vein.
Tell us what you would do
How would you manage this patient? Submit your comments below.