Backgrounder: recognizing and diagnosing primary HIV infection. (Backgrounder).

"If you want a bottom line, it's that we should all be actively looking for patients in the acute stages of HIV infection and either treating them or getting them into studies."

Eric Rosenberg, MD
Harvard Medical School
The PRN Notebook, June 2001

Introduction. Researchers have expended much time and effort studying primary HIV infection (PHI). Large amounts of money have been spent identifying patients in the earliest stage of HIV infection, studying their immune responses, and conducting studies of therapeutic interventions. Given that the entire course of HIV disease spans years as opposed to the weeks or months of PHI, it is logical to ask, what is the urgency of identifying persons experiencing PHI? Why intervene with aggressive antiviral therapy given that most symptoms associated with PHI are self-limited and that current trends in treatment are leaning towards waiting longer before initiating therapy? Why study immune responses during this brief period of acute illness?

There are no definitive answers to these questions, but scientific evidence thus far suggests there are some good reasons for exploring PHI. Theoretical rationale cited for identifying, treating, and studying patients with PHI have generally been described as follows:

* Treating patients during PHI may preserve critical immune function and later allow patients to control viremia without therapy.

* Treating patients during PHI may alter the initial viral "set point," which may ultimately affect the rate of disease progression.

* Treating patients during PHI may lower the rate of viral diversification.

* Treating patients during PHI may reduce the severity of the acute retroviral syndrome.

* Persons with PHI may be highly infectious. Some epidemiological models suggest that over half of HIV transmission occurs during PHI. (1) Identifying PHI patients gives health care providers an opportunity to counsel them about their risk behaviors and treat them to decrease their infectiousness.

* Identifying patients with PHI may provide information about the transmission rate of resistant virus and resistance testing during PHI may help determine the optimal treatment for patients once the decision to begin therapy is made.

While there was hope early during the development of highly active antiretroviral therapy (HAART) that treatment during PHI would decrease viral seeding of lymphoid tissue and facilitate eradication, this strategy is no longer feasible. It is now known that viral reservoirs are established in resting memory CD4 T cells early in infection and certainly before the development of HIV-specific antibodies.

The initial host-virus interaction during PHI represents a unique window of opportunity for intervention and for studying HIV immunopathogenesis. Many researchers speculate that such study will ultimately lead to a better understanding of the parameters of effective immune control of HIV and eventually result in the development of an effective vaccine or other method of immune-mediated viral containment.

PHI in context. The natural history of HIV disease is generally classified into the following stages:

* Viral transmission

* Primary HIV infection (with or without symptoms)

* Seroconversion (evidenced by the detectability of HIV-specific antibodies)

* Asymptomatic chronic infection

* Symptomatic HIV infection

* AIDS (defined as the occurrence of an indicator condition, such as an opportunistic infection, or a CD4 T cell count less than 200 cells/[mm.sup.3]) and

* Advanced HIV infection (characterized by a CD4 T cell count less than 50 cells/[mm.sup.3]).

As noted earlier, the course of PHI is limited to a few weeks or months, whereas the entire course of HIV infection can span many years.

Specifically, PHI is the period after infection with HIV but before the development of detectable antibodies. (See Figure 1.) It is a period of active HIV replication and transient immune suppression. Viremia rises rapidly in the plasma, often reaching levels in excess of 1 million copies per milliliter, with widespread dissemination into lymphoid organs throughout the body. While some patients have no clinical symptoms during PHI, it has been estimated that 50% to 90% of patients are symptomatic with an illness resembling infectious mononucleosis or influenza and that a large percentage of these patients present for care at health care facilities. (2) This symptomatic manifestation of infection is generally referred to as "acute retroviral syndrome" (ARS). Some of the most prevalent symptoms associated with ARS are listed in Table 1.

[FIGURE 1 OMITTED]

During the short course of PHI, the occurrence of ARS coincides with the increase in plasma viremia, which as noted earlier can reach high levels. Symptoms of ARS are estimated to begin 2 to 6 weeks after infection and resolve in approximately 14 days in most cases. (2,3) The resolution of symptoms coincides with the decrease in plasma viremia that is associated with the emergence of HIV-specific CD8 T cells and the later emergence of HIV-specific antibodies. (4,5) Plasma viremia can drop approximately 2 to 3 logs as the immune system begins to respond to the infection. After fluctuating for some time, viremia then appears to stabilize at a "set point."

The exact determinants of the viral set point are not known, but factors such as the immune response of the host, the number of available target CD4 T cells, the host's degree of immune activation, the extent of trapping and sequestration of HIV and infected CD4 T cells in the germinal centers of lymphoid tissue, as well as the replicative capacity of the viral strain, may contribute to the set point. (6,7) The set point generally predicts the rate of disease progression, suggesting that this initial viral-host interaction during PHI is critical in HIV immunopathogenesis.

Who is suspect for PHI? The problem with identifying PHI patients is two-fold. First, the symptoms associated with ARS are nonspecific and can mimic other common illnesses. Second, there is no single diagnostic test that is cost-effective and sufficiently specific and sensitive enough to conduct broad-based testing of the suspected patient population.

Studies have been conducted to determine the usefulness of symptoms and virologic tests for diagnosing PHI. Daar et al. recently published a prospective cohort study of 436 patients with potential exposure to HIV who reported symptoms compatible with PHI. (8) In this study patients were referred from clinics, testing centers, emergency departments, and community physicians to 2 research hospitals in Los Angeles and San Diego. The patients were divided up by institution and time into 3 cohorts as outlined in Table 2. The demographics of the 3 cohorts were similar. Overall, 89% of the study patients were male, also 74% were white, 13% Hispanic, and 9% African American. Seventy-seven percent of the patients were gay men, 18% heterosexual women, and 4% intravenous drug users.

All 3 cohorts had virologic testing but only the 255 patients in Cohort 2 were questioned about specific symptoms. The patients in Cohort 2 were divided into those with PHI versus those who were uninfected or who had chronic HIV infection. PHI was defined as a positive HIV RNA (viral load) and a negative antibody test, or an indeterminate Western blot and negative antibody test in the preceding 12 weeks. Of the total, 40 patients were confirmed to be experiencing PHI and 164 were either uninfected or chronically infected.

PHI patients were more likely to be gay and exposed to an HIV-infected person. A comparison of symptoms showed that PHI patients were more likely to experience fever, myalgia, arthralgia, rash, or night sweats. Only these symptoms and the lack of nasal congestion turned out to be statistically significant predictors of PHI. (See Table 3 on next page.) Combining fever, myalgia, and rash increased the predictive value of these symptoms. However, no combination of symptoms identified more than 75% of PHI patients.

Despite the increased incidence of certain clinical symptoms in PHI patients, the authors conclude that no symptoms have sufficient sensitivity or specificity for PHI to allow for targeted screening of at-risk patients. Thus, it may be necessary to establish a low threshold of symptoms and test many seronegative patients in order to maximize the number of PHI patients who are properly identified.

In contrast, Frederick M. Hecht, MD, co-director of the Options Project at the University of California, San Francisco has reported some slightly different results. (9) The Options Project, which has been in existence since 1996, has screened 444 patients for PHI and early HIV infection. (Early infection is defined as infection and subsequent seroconversion within 12 months of entering the study.) Hecht has been quoted as stating that "[T]he key symptoms [are] rash and fevers, followed by loss of appetite, arthralgias, and pharyngitis." Using logistic-regression models, this group has determined that the odds of rash predicting PHI are 3.7 (p=0.002) and the odds of fever predicting PHI are 3.4 (p=0.009). According to Hecht, "In other viral infections, we don't often see fevers and we rarely see rash. These two symptoms, especially if accompanied by some of the other symptoms frequently seen in PHI, should heighten the level of suspicion." (9)

Diagnosis of PHI. During PHI there is a "window period" of approximately 2 to 3 months after initial infection before HIV-specific antibodies in the blood can be detected by the standard antibody test (ELISA). During this window period other tests must be used to diagnose PHI. The most common tests used to diagnose PHI are the quantitative HIV RNA (viral load), which uses polymerase chain reaction to detect the presence of HIV, and the p24 antigen test, which detects a viral protein and indicates ongoing viral replication.

In an ideal world a screening test for PHI would have 100% sensitivity. A test with that level of sensitivity would detect everyone who has been infected with the virus (i.e. no one who was positive would go undetected). The perfect confirmation test would also be one that has 100% specificity (i.e. there would be no false-positive results). It is important to note that, in the context of PHI, testing for HIV antibodies is irrelevant since by definition PHI patients have not been infected long enough to develop HIV-specific antibodies.

Unfortunately, current tests used in PHI are neither 100% sensitive nor specific. In the study cited above, Daar et al. (8) tested patients in the 3 cohorts for p24 antigen and HIV RNA. Because there is no current gold standard test for HIV infection, the authors assumed that the HIV RNA test was 100% sensitive. The results showed that while the p24 antigen test has a specificity of 100% (there were no false positives), it only has a sensitivity of 88.7% (5 of the 40 PHI patients were not detected by the test). Also, while the researchers assumed that the HIV RNA test had a sensitivity of 100%, the study results demonstrated that the test had a specificity of 97.4% (8 of the 303 patients tested had false-positive results as confirmed by repeat testing). A more detailed summary of the results is listed in Table 4. Because the specificity of the HIV RNA test is not 100%, most experts recommend that results with low viral titer (less than 10,000 copies/mL) be approached with caution since they may be false positives.

The problems with these tests are further complicated by their relative costs. While the p24 antigen test is cheap and easy to access, studies have demonstrated that it may fail to detect up to 20% to 25% of PHI patients. In fact, Christopher D. Pilcher, MD, of the University of North Carolina has determined that p24 positivity is time dependent; it may be necessary to test within 3 weeks of the onset of PHI symptoms to get an accurate reading. (10) After 3 weeks, up to a quarter of PHI patients will have negative p24 antigen test results. (See Figure 3.) On the other hand, while the HIV RNA test is highly sensitive, it is a relatively expensive test that may still result in some false-positive readings.

[FIGURE 3 OMITTED]

The costs and benefits of using these tests in the context of PHI is best illustrated by a theoretical scenario outlined in an editorial accompanying the Daar et al. article cited above. (11) In that editorial the authors posit that an urban emergency department might have 50,000 patient visits in a year. Of those, it is reasonable to assume that 500 patients will be screened for PHI based on symptom presentation. If 1% of these patients (5 patients) had PHI, a p24 antigen test would be expected to diagnose 4 of the 5 patients. On the other hand, the use of an HIV RNA test would diagnose all 5 patients but up to 3% (15 patients) might have false-positive results. These false positives would require extensive counseling and further testing. If the cost of a p24 antigen test was $75 per test, then the cost of screening those 500 patients would be $37,500. By contrast, if the cost of providing HIV RNA testing was $220 per test, the cost of testing all 500 patients would be $110,000.

Obviously, the choice of tests will depend on the circumstances. For general screening, the HIV RNA may not be cost effective and a p24 antigen test may be more appropriate. However, when patients exhibit symptoms and behavior that are highly consistent with HIV infection, an HIV RNA test may be merited.

In addition to the virologic tests described above, the differential diagnosis for presumptive PHI includes viral hepatitis, mononucleosis, herpes virus infection, anthrax infection, and secondary syphilis. A complete blood cell count usually reveals lymphopenia, thrombocytopenia, and atypical lymphocytes.

Patients suspected of having PHI should be screened for other sexually transmitted diseases, including syphilis and hepatitis. Conversely, patients with other STDs should be prospectively identified as high-risk for PHI.

Conclusions. Recognizing PHI in a patient remains a tricky business. There is some evidence to indicate that PHI symptoms may be nonspecific. However, some researchers point to the presence of rash and fever as being particularly predictive of PHI. Taking a history of a patient's risk behaviors including unprotected sexual encounters and sharing of needles is critically important. Physicians should be aware that patients are often reluctant to disclose that type of information. A physician therefore must maintain a higher level of suspicion for HIV infection. Clinicians should remember that p24 antigen tests, while inexpensive and easily accessed, may miss a large percent of PHI patients. There is evidence that the accuracy of p24 antigen tests results may be time dependent. Patients tested after 3 weeks of the onset of PHI symptoms are likely to be missed by p24 antigen tests. In cases where risk behavior and symptoms indicate a strong possibility of PHI, an HIV RNA test may be merited. However, treating physicians should remember that some false positives will result and HIV RNA results with low copy numbers are especially suspect for being false positives.

Although treatment guidelines promulgated by the US Department of Health and Human Services recommend that clinicians offer antiretroviral therapy to patients with PHI, (12) treating this early stage of infection remains controversial. The toxicities of anti-HIV drugs are now widely known and the guidelines provide no direction on the duration of treatment for PHI. Consequently, some clinicians are reluctant to offer a hazardous, open-ended course of therapy to patients who will suffer no ill effects from HIV infection for years.

Still, properly recognizing and diagnosing PHI has merits, not the least of which is accurately informing patients of the nature of their illness and treating it in a way appropriate to the circumstances. Other compelling theoretical rationale, discussed in greater detail in other articles in this issue, argue for at least enrolling patients with PHI in clinical trials where available, and perhaps for treating them. Time will tell if the effort, time, and money spent on identifying, diagnosing, and studying PHI patients will have been worthwhile. However, ignoring this phase of HIV infection will only insure that critical questions about transmission and pathogenesis might never be fully answered.

TABLE 1. Signs and Symptoms
Associated with Acute
Retroviral Syndrome

Signs and symptoms      Frequency

Fever                   96%
Lymphadenopathy         74%
Pharyngitis             70%
Rash                    70%
Myalgia or arthralgia   54%
Diarrhea                32%
Headache                32%
Nausea and vomiting     27%
Hepatosplenomegaly      14%
Weight loss             13%
Thrush                  12%
Neurologic symptoms     12%

From the US Dept. of Health and Human Services.
Guidelines for the Use of Antiretrovirals in Adults and
Adolescents. August 2001.
TABLE 2. Serologic and Virologic Results in
Patients Screened for PHI

COHORTS

* Cohort 1:127 patients 3/93-8/95 (Los Angeles)
* Cohort 2:255 patients 6/96-7/99 (Los Angeles)
* Cohort 3:54 patients 6/96-3/00 (San Diego)
  Total patients: 436

Virologic testing in all 3 cohorts
Symptoms by questionnaire for cohort 2 only

Adapted from Daar ES, et al. (8)
TABLE 3. Clinical Predictors of PHI in Cohort 2

Sign or symptom    PHI (n = 40)   Not PHI (n = 164)

Fever              35 (87.5%) *   82 (50%)
Myalgias           24 (60.0%) *   43 (26.2%)
Rash               23 (57.5%) *   34 (20.7%)
Night sweats       20 (50%) *     52 (31.7%)
Nasal congestion   7 (17.5%)      62 (37.8%) *

* P < 0.05

Note: malaise, sore throat, lymphadenopathy, and
thrush were among the symptoms that were not statistically
significant predictors of PHI.

Adapted from Daar ES, et al. (8)
TABLE 4. Serologic and Virologic Results in Patients Screened for PHI

Variable total (n = 436)    Positive HIV RNA    Positive p24 antigen

Primary HIV
  * Negative HIV antibody   41/41 (100%)        35/40 (87.5%)[73-95.8]
  * Indeterminate Western
      blot                  13/13 (100%)        12/13 (92.3%) [64-99.8]

Chronic HIV                 60/60 (100%)        10/55 (18.2%)
                            [not performed in   [not performed
                            19 patients]        in 24 patients]

Uninfected                  8/303 (2.6%)        0/303 (0%)

Numbers in square brackets are 95% confidence intervals. No confidence
intervals are given for HIV RNA since sensitivity was assumed to be
100%.

Adapted from Daar ES, et al. (8)

REFERENCES

(1.) Jacquez JA, Koopman JS, Simon CP. Longini IM Jr. Role of the primary infection in epidemics of HIV infection in gay cohorts. J Acquir Immune Defic Syndr. 1994;7(11):1169-1184.

(2.) Schacker T, Collier AC, Hughes J, et al. Clinical and epidemiologic features of primary HIV infection. Ann Intern Med. 1996;125(4):257-264.

(3.) Daar ES. Virology and immunology of acute HIV type 1 infection. AIDS Res Hum Retroviruses. 1998; 14(Suppl 3):S229-234.

(4.) Fauci AS, Pantaleo G, Stanley S, Weissman D. Immunopathogenic mechanisms of HIV infection. Ann Intern Med. 1996;124(7):654-663.

(5.) Koup RA, Safrit JT, Cao Y, et al. Temporal association of cellular immune responses with the initial control of viremia in primary human immunodeficiency virus type 1 syndrome. J Virol. 1994;68(7):4650-4655.

(6.) Mellors JW, Rinaldo CR Jr, Gupta P, et al. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science. 1996;272(5265):1167-1170.

(7.) Quinn TC. Acute primary HIV infection. JAMA. 1997;278(1):58-62.

(8.) Daar ES, Little S, Pitt J, et al. Diagnosis of primary HIV-1 infection. Los Angeles County Primary HIV Infection Recruitment Network. Ann Intern Med. 2001;134(1):25-29.

(9.) Hecht, FM. Recent issues in primary HIV: from diagnosis to oral sex transmission. The PRN Notebook. 2000;5(4):9-12.

(10.) Koopman J, Pilcher CD. Where epidemiology meets biology: primary HIV infection and sexual transmission. The PRN Notebook. 2001;6(4):20-27.

(11.) Flanigan T, Tashima KT [editorial]. Diagnosis of acute HIV infection: it's time to get moving! Ann Intern Med. 2001;134(1):75-77.

(12.) US Dept. of Health and Human Services. Guidelines for the Use of Antiretroviral Agents in HIV-Infected Adults and Adolescents. August 2001. http://www.hivatis.org/trtgdlns.html#Adult