Polymerase chain reaction (PCR) procedures to detect HIV DNA in blood samples are increasingly being used for early diagnosis of HIV infection in infants who are born to HIV-positive mothers. However, PCR methods appear to be appreciably less sensitive in the early days of life than later in infancy (1-3), presumably because much lower levels of HIV DNA are present in blood at or near birth than later in infancy. Interestingly, in a study by Comeau et al., vertically infected infants who developed rapidly progressive HIV disease seemed especially likely to have blood specimens that were positive by DNA PCR within a few days of birth (4), which suggests that such infants have relatively high blood levels of HIV DNA. In contrast, De Rossi et al. (5) found that detection of HIV DNA by age 15 days did not closely correlate to the development of severe disease; however, at 4-8 weeks of age, seven infants who developed AIDS by 12 months of age had higher levels of HIV DNA per 105 peripheral blood mononuclear cells (PBMC) (360-947 copies) than 11 infants (with 10-314 copies) who did not develop AIDS. Others have noted that seriously ill children and adults tend to have higher blood levels of HIV and such components as its antigen than do those less seriously ill (6-11). Furthermore, in adults, the amount of HIV DNA in PBMC and mRNA in plasma and PBMC appears to predict the subsequent development of HIV disease (12-14).
Our aims in the present study were (a) to characterize changes in HIV DNA load in peripheral blood mononuclear cells (PBMC) of infected infants and young children as they age, (b) to better understand the reasons for the apparent insensitivity of HIV PCR early in infancy, and (c) to determine whether HIV DNA load in early infancy predicts disease progression. To accomplish these aims, we performed endpoint dilution PCR tests to quantitate HIV DNA levels in PBMC specimens from infected infants and young children, particularly those who were initially sampled prior to 6 months of age and whose illness status at 24 months of age is now known. In addition, we similarly tested PBMC specimens from HIV-infected mothers about the time of the birth of their infants, in order to compare HIV DNA levels in infants and children with those of adults.
Our pediatric study patients were newborns, infants, or young children when initially sampled. All were born to HIV-positive mothers and eventually were determined by Centers for Disease Control criteria (15) to be HIV-infected. Commonly, these patients were examined and blood specimens were obtained for PCR studies at intervals of 3 months or less. One group of mothers consented to be study subjects themselves. These mothers were asymptomatic at the time of recruitment. The study began in January 1989.
In efforts to relate HIV DNA levels to disease progression, our pediatric patients were categorized on the basis of their most serious clinical findings by age 24 months. Severely ill infants and children were defined as those who developed encephalopathy, a serious AIDS-defining opportunistic infection—primarily Pneumocystis carinii pneumonia (PCP)—or had died of AIDS, sepsis, or sudden infant death syndrome (SIDS) by age 24 months. Moderately ill patients were those with less severe infections, failure to thrive, developmental delay, or lymphoproliferative interstitial pneumonia (LIP). Mildly ill patients were asymptomatic or had only nonspecific findings. For some analyses, the moderately ill and mildly ill pediatric patients were grouped together and categorized as not severely ill. Our study was approved by the Institutional Review Board at Children's National Medical Center and several nearby hospitals where infants were delivered.
The PCR methods have been described in some detail (16,17). Briefly, our testing for HIV proviral DNA was performed with coded PBMC lysates and 40 cycles of amplification, using gag primers SK 38/39 (with biotinylated probe SK 19) and SK 145/150 (with biotinylated probe SK 102). Individual PCR tests were performed with a PBMC lysate obtained from a measured volume of the original blood sample. Specimens found positive for HIV DNA were titrated (typically using five-fold or two-fold serial dilutions of PBMC lysates in water near the endpoint of PCR reactivity) to determine the highest specimen dilution in which HIV DNA could be detected by PCR with each primer pair. HIV DNA copy numbers were then calculated to reflect the reciprocal of the highest dilution of 0.1-ml volumes of original blood samples in which HIV DNA was detected with the most sensitive primer pair. For some analyses, the HIV copy number per 105 CD4 cells was also calculated by dividing the HIV copy number per 0.1 ml of blood by the CD4 cell count per cubic millimeter of blood.
Age-Related Variations in HIV DNA Levels
Blood levels of HIV DNA in our vertically infected pediatric patients varied over a wide range, but in individual patients they typically followed a characteristic age-related pattern: continually increasing with increasing age to a peak between 4 and 8 months, and thereafter rather steadily declining. Median HIV DNA copy numbers for all tested pediatric samples increased from 77 copies per 0.1 ml blood sample at birth, to 530 at age 1 to 2 months, and peaked at 2,700 at age 6 to 8 months (Table 1). Thus, the median HIV DNA copy number per 0.1 ml was 6.9-fold higher in the second month of life than at birth, and 35-fold higher at age 6 to 8 months than at birth. Median titers then tended to decrease with increasing age and were especially low (63/0.1 ml) in samples obtained after age 3 years.
The proportion of pediatric samples with ≥1,000 HIV DNA copies per 0.1 ml of blood followed a similar pattern, rising from 25.8% in the first month of life to 71.4% at age 6 to 8 months, and thereafter declining to 18.5% after 3 years of age (Table 1). The decline in blood levels of HIV DNA nearly always began prior to the use of AZT or other antiretroviral therapy.
HIV Copies in Mothers
Forty-nine infected mothers were sampled near the time of the birth of their infants. These mothers had a median HIV copy number of 22/0.1 ml, considerably lower than the median level for any group of pediatric patients, and only 5 (10.2%) had ≥1,000 HIV DNA copies in their 0.1-ml sample (Table 1). The eight mothers who are known to have transmitted HIV to their infants tended to have higher HIV DNA copy numbers (median of 198) than 17 mothers who definitely did not transmit infection (median of 27 copies; p = 0.0329 using the Mann-Whitney rank-sum test).
As a further indication of the relatively low blood levels of HIV DNA in the infected mothers, we should also note that six (12.2%) of the 49 tested mothers had such low levels of HIV DNA in their PBMCs that their specimens consistently tested negative in our highly sensitive diagnostic PCR tests and retests for HIV DNA. No infected mothers with consistently negative PCR tests are known to have transmitted HIV to their infants. Among the mothers with CD4 counts measured near delivery, five who are known to have transmitted HIV to their infants had a higher median level of HIV DNA per 105 CD4 cells (425 copies) than 15 who definitely did not transmit (24 copies).
HIV DNA Copies and Illness Severity
In each of nine different age groups from early infancy into early childhood, infected pediatric patients who became severely ill by age 24 months were more likely to have at least 1,000 HIV DNA copies per 0.1 ml blood sample than were infected patients who did not become severely ill (Table 2). In addition, peak HIV DNA titers, as evidenced by the highest proportion of patients with ≥1,000 copies, and by the highest median copy numbers in tested specimens, occurred at an earlier age for those who became seriously ill than for those less seriously ill (4-6 months vs. 6-8 months).
Ages When DNA Levels Peaked
To better establish that there were differences in the ages at which HIV DNA blood levels peaked in pediatric patients who did become severely ill versus those who did not, we determined the age at which peak HIV DNA titers per 0.1 ml of blood occurred in the first 24 months of life for 39 patients who had at least two titrated samples in their first 8 months of life. The mean age at titer peak for 15 who became severely ill was significantly earlier: 4.4 months (132.9 ± 101.1 days) versus 7.8 months (235.5 ± 168.0 days) for 24 who were not severely ill (p = 0.0224 by Student's t test). Thus, the HIV DNA copy number peak in those who became severely ill occurred 3.4 months earlier than in those who were not severely ill.
Titers in Five Illness Categories
Tables 3 and 4 provide an especially clear indication that HIV copy numbers early in life could be used to predict the relative severity of HIV disease that will develop by age 2 years. Whether presented in terms of copy numbers per 0.1 ml of blood or per 105 CD4 cells, both the proportion of patients and specimens with at least 1,000 HIV DNA copies, and the median specimen titers, were highest and similar for those who died or became severely ill, were of intermediate level for those who were moderately ill or had LIP, and were lowest for specimens from those who were minimally ill. For example, ≥1,000 HIV DNA copies were present per 0.1 ml of blood from 23 (85.9%) of 27 patients who either died or became severely ill, and from 13 (52.0%) of 25 who were of intermediate illness status, but in only 2 (22.2%) of nine patients who were mildly ill (chisquare = 13.3, p ≤ 0.001) (Table 3). Furthermore, median copy numbers per 0.1 ml specimens from those who died or became severely ill (median = 5,026) were approximately seven-fold higher than from those of intermediate illness status, and well over 100-fold higher than from those who were only mildly ill. Per 105 CD4 cells (Table 4), ≥1,000 HIV DNA copies were present prior to 6 months of age in 16 (80.0%) of 20 patients who died or became severely ill, in 10 (43.5%) of 23 patients who were moderately ill or who had LIP, but in only 1 (11.1) of nine patients who were mildly ill (chi-square = 13.0, p = 0.002). In the various illness groups, median copy numbers per 105 CD4 cells were similar to those previously seen per 0.1 ml of blood and were markedly increased with increasing illness severity.
Infected Patients Who Died
Table 5 presents additional findings relative to nine infected study children who died of presumably HIV-related disease. Six (67%) of these patients died between 3.0 and 4.2 months of age, and eight (89%) died in the first year of life. Of six who died with AIDS or sepsis, five (83%) had >3,000 copies of HIV DNA present in a 0.1-ml blood sample prior to 4 months of age. Three patients apparently were SIDs victims (a very high incidence of SIDS, given our study size). Prior to 5 months of age, two of the SIDS victims had >6,000 copies of HIV DNA present per 0.1 ml of blood, which strongly suggests that HIV infection was causally associated with their death. In addition, >1,100 HIV DNA copies were present per 105 CD4 cells in each of the four dead patients who were appropriately tested prior to age 5 months (Table 5).
Patients with Minimal Disease Progression
Over a wide range of ages, our vertically infected pediatric patients who remained minimally ill beyond 2 years of age tended to have very low blood levels of HIV DNA. Findings from two such patients especially illustrate this point. The first child was still only mildly ill at 5 years of age. A specimen for PCR testing was obtained from this patient every 3 to 6 months between ages 62 days and 37½ months. In none of his nine specimens was more than eight copies of HIV DNA detected per 0.1 ml, or more than 1.5 copies per 105 CD4 cells. Furthermore, in three of his specimens (at 16, 22, and 37½ months of age) repeated testing failed to detect any HIV DNA. Another child with minimal disease progression was sampled five times between 39 days and 19 months of age. In none of these specimens was more than 6 HIV DNA copies detected per 0.1 ml of blood, or 3 copies per 105 CD4 cells. HIV infections in such patients obviously could be missed by insensitive diagnostic PCR tests.
Earliest Prediction of HIV Severity
Five infected patients whose illness status is now known were tested for HIV DNA levels within 48 hours of birth. Four of these patients were not severely ill at age 24 months. On the day of birth their respective levels of HIV DNA per 0.1 ml of blood were: 0, 2, 77, and 150 copies. The remaining patient became severely ill during infancy; at age 2 days she had 2 HIV DNA copies per 0.1 ml of blood. These findings suggest that within 48 h of birth, neither the mere presence of HIV DNA nor the presence of easily detectable levels of HIV DNA is a reliable predictor of the development of severe HIV disease.
All of these patients subsequently were sampled at least once during the first 6 months of life. The HIV DNA copy numbers per 0.1 ml of blood in those who did not become severely ill were, respectively, 538 on day 31; 18 and 819 on days 25 and 123, 540 on day 36; 120 on day 27 and 216 on day 53. The patient who became severely ill had copy numbers of 7,700 on day 35, 27,000 on day 99, and 13,500 on day 139. Thus, for these patients, HIV DNA titers consistent with their eventual clinical outcome were present by about 1 month of age.
Few of our patients whose serious illness outcome is now confirmed were sampled between 3 and 30 days of age. Therefore, it is not clear from our data how soon before 1 month of age one might make a reliable prediction of HIV disease severity.
Predicting Severe HIV Disease
Upon retrospective analysis, the detected presence of certain age-related levels of HIV DNA per 0.1 ml of blood—>800 copies between ages 3 weeks and 2 months, >1,000 copies from ages 2 to 4 months, and >2,500 copies from ages 4 to 6 months—would have maximized our identification of patients who became severely ill by age 24 months while minimizing incorrect identification. Had such levels been the indicators for identifying infants in Table 2 who were 3 weeks to 2 months of age, seven (63.6%) of 11 patients who became severely ill as compared with four (22.2%) of 18 non-severely ill would have tested positive (chi-square = 4.98, p = 0.026). Furthermore, between 3 weeks and 6 months age, 30 (76.9%) of 39 who became severely ill compared with 14 (28.0%) of 50 non-severely ill patients had such copy numbers (chi-square = 20.98, p ≤ 0.0001).
Among patients who were not severely ill, it remains to be seen whether the progression of HIV disease after 24 months of age will tend to be more rapid in those who had especially high blood levels of HIV DNA. However, in view of the tendency over a wide range of patient ages for higher titers to be present in those who became severely ill (Table 2), it seems likely that this will occur.
It is evident from the present study that copy numbers of HIV DNA in 0.1-ml blood samples from our vertically infected pediatric patients characteristically increased with increasing age to a peak between 4 and 8 months of age, and then rather steadily declined. Although our data are not shown, essentially the same pattern of age-related increase, peak, and fall in copy number was seen when measured in terms of copies of HIV detected per 105 CD4 cells.
It has been suggested that infants with any detectable HIV DNA in their blood within 48 h of birth are those who were infected prior to labor and delivery (18), and that infants who become infected during gestation may have a more extensive infection (presumably with more HIV DNA copies in their blood) and more accelerated disease than those who become infected during labor and delivery (11). The data of Comeau et al. (4) suggest that illness-predictive differences in blood levels of HIV DNA may exist in the first week of life. In contrast, De Rossi et al. (5) found such differences only after 4 weeks of age. Five of our infected infants whose illness status at 24 months of age is now known were tested for HIV DNA within 48 h of birth. In them, neither the presence nor the level of HIV DNA at that time was a reliable indicator of eventual disease progression. However, our infants with the highest blood levels of HIV DNA ≥3 weeks after birth were especially prone to develop PCP and/or encephalopathy or to die at an early age. Our findings relative to patients with LIP are consistent with observations by others (19,20) that LIP is a less severe manifestation of pediatric AIDS than is PCP or encephalopathy. Also, in keeping with data from HIV-infected adults who remained in unusually good health for long periods of time (21,22), our infected pediatric patients with the least progressive HIV disease tended to maintian very low blood levels of HIV DNA.
The low HIV DNA blood levels in most of our study mothers during labor is consistent with their generally good health at that time. In addition, the tendency for higher HIV DNA levels to be present in mothers who transmitted infection agrees with similar findings of others (23).
In our study, the decline in blood levels of HIV DNA that regularly begins during the first year of life tended to occur first in those who became severely ill. In them, a more rapid depletion of cells capable of producing HIV would have contribued somewhat to the earlier decline. However, since above-average blood levels of HIV DNA tended to be present well after infancy in those who became severely ill, the effect of cell depletion alone does not fully explain the observed differences. Our severely ill patients also tended to receive AZT or other antiretroviral therapy at an earlier age than did the less severely ill. However, the fall in HIV DNA titers typically began prior to the use of antiretroviral therapy; thus, treatment was not a key initiator of the titer decline.
In recently infected adults, a rapid decline in cell-free and cell-associated HIV in blood occurs at seroconversion, indicating that a host's immune response can markedly limit HIV replication (24,25). Accordingly, it would seem reasonable that the timing of titer declines in our infants was importantly influenced by an immune response, possibly as follows: more virus, and thus more immunogenic mass, would tend to be present at an early age in those who were to become severely ill. Levels of antibody and/or cell-mediated immunity that would noticeably inhibit virus production thus would tend to be present earlier in those who were to become the most severely ill.
Several findings suggest that half to two-thirds of the vertical infections with HIV occur during labor and delivery (3,26). If so, one might expect that infants with intrapartum infections would have age-related levels of HIV DNA in their PBMCs that are different from those who were infected some months earlier. By analogy to findings for virus and virus component levels in recently infected adults (24,25), pediatric patients infected during labor and delivery might be expected to show a sharp burst in HIV DNA very shortly after birth, followed by an extended period of lower blood levels of the DNA. Only much later would there be a rather steadily rising level of HIV DNA. This definitely is not the pattern seen in our patients, but it is not clear what—if anything—the observed pattern indicates about the timing of transmission. Such findings could be an indication that most HIV transmission to infants actually occurs well before labor and delivery. Conversely, the absence of a clearly evident sharp peak in HIV DNA levels in the first month or two after birth may merely result from the inhibitory effect of maternally derived antibody and/or the slow development of immune responses by an immature immune system.
That there apparently is a much lower blood level of HIV DNA in virtually all vertically infected infants near birth than several months later is quite understandably a major reason why various DNA PCR methods have often detected no more than 30-50% of maternally derived HIV infections during the first week after delivery and only 70% of such infections during the first month of life (1,2).
In part because of our more frequent finding of HIV DNA in blood specimens of young infants (in eight (89%) of nine infected infants who were sampled within 48 h of birth, and in 18 (95%) of 19 infected infants sampled in the first month of life) (17), it appears that suboptimal PCR procedures have also contributed to failures of PCR testing early in life. Relative to this point, as simple a methodologic change as merely reamplifying apparently negative diagnostic lysates after diluting them five-to 10-fold (thus removing the effect of PCR inhibitors) often considerably improved our own detection sensitivity. It has also done so in at least one other laboratory of which we are aware, and likely would do so for others.
Acknowledgment: These studies were supported in part by a grant from the Alexander and Margaret Stewart Trust and by the Discovery Fund of Children's National Medical Center, Washington, D.C. We thank Sheila A. Suckling for her considerable help in the preparation of this manuscript.
1. Albert J, Biberfeld G, Borkowsky W. Report of a consensus workshop, Siena, Italy, January 17-18, 1992. J Acquir Immune Defic Syndr
2. Comeau AM, Harris J-A, McIntosh K, Weiblen BJ, Joff R, Grady GF. Polymerase chain reaction in detecting HIV infection among seropositive infants
: relation to clinical status and age to results of other assays. J Acquire Immune Defic Syndr
3. Dunn DT, Brandt CD, Krivine A, et al. The sensitivity of HIV-1 DNA polymerase chain reaction in the neonatal period and the relative contributions of intra-uterine and intrapartum transmission. AIDS
4. Comeau AM, Hsu H-W, Schwerzler M, Mushinsky G, Grady GF. Detection of HIV in specimens from newborn screening programs. N Engl J Med
5. De Rossi A, Giaquinto C, Ometto L, et al. Replication and tropism of human immunodeficiency virus type 1 as predictors of disease outcome in infants
with vertically acquired infection. J Pediatr
6. Alimenti A, O'Neill M, Sullivan JL, Luzuriaga K. Diagnosis of vertical human immunodeficiency virus type 1 infection by whole blood culture. J Infect Dis
7. Luzuriaga K, McQuilken P, Alimenti A, Somasundran M, Hesselton RA, Sullivan JL. Early viremia and immune responses in vertical human immunodeficiency virus type 1 infection. J Infect Dis
8. Saag MS, Crain MJ, Decker WD, et al. High-level viremia in adults and children infected with human immunodeficiency virus: relation to disease state and CD4+ lymphocyte levels. J Infect Dis
9. Borkowsky W, Krasinski K, Paul D, et al. Human immunodeficiency virus type 1 antigenemia in children. J Pediatr
10. Dickover RE, Dillon M, Gillette SG, et al. Rapid increases in load of human immunodeficiency virus correlate with early disease progression and loss of CD4 cells in vertically infected infants
. J Infect Dis
11. Pizzo PA, Wilfert CM. Report of a consensus workshop, Siena, Italy, June 4-6, 1993: Markers and determinants of disease progression in children with HIV infants
. J Acquir Immune Defic Syndr
12. Chevret S, Kirstetter M, Mariotti M, Lefrère F, Frottier J, Lefrère J-J. Provirus copy number to predict disease progression in asymptomatic human immunodeficiency virus type 1 infection. J Infect Dis
13. Saksela K, Stevens C, Rubenstein P, Baltimore D. Human immunodeficiency virus type 1 mRNA expression in peripheral blood cells predicts disease progression independently of the numbers of CD4+ lymphocytes. Proc Natl Acad Sci
14. Verhofstede C, Reniers S, Wanzeele FV, Plum J. Evaluation of proviral copy number and plasma RNA level as early indicators of progression in HIV-1 infection: correlation with virological and immunological markers of disease. AIDS
15. Centers for Disease Control. Classification system for human immunodeficiency virus (HIV) infection in children under 13 years of age. MMWR
16. Brandt CD, Rakusan TA, Sison AV, et al. Detection of human immunodeficiency virus type 1 infection in young pediatric patients by using polymerase chain reaction and biotinylated probes. J Clin Microbiol
17. Brandt CD, Rakusan TA, Sison AV, Saxena ES, Ellaurie M, Sever JL. Human immunodeficiency virus infection in infants
during the first 2 months of life. Arch Pediatr Adolesc Med
18. Bryson YJ, Luzuriaga K, Sullivan JL, Wara DW. Proposed definitions in utero versus intrapartum transmission of HIV-1. N Engl J Med
19. Thomas P, Singh T, Williams R, Blum S. Trends in survival for children reported with maternally transmitted acquired immunodeficiency syndrome in New York City, 1982 to 1989. Pediatr Infect Dis J
20. Blanche S, Tardieu M, Duliege A-M, et al. Longitudinal study of 94 symptomatic infants
with perinatally acquired human immunodeficiency virus infection: evidence for a bimodal expression of clinical and biological symptoms. Am J Dis Child
21. Pantaleo G, Menzo S, Vaccarezza M, et al. Studies on subjects with long-term nonprogressive human immunodeficiency virus infection. N Engl J Med
22. Cao Y, Qin L, Zhang L, Safrit J, Ho D. Virologic and immunologic characterization of long-term survivors of human immunodeficiency virus type 1 infection. N Engl J Med
23. Borkowsky W, Krasinski K, Caoy, et al. Correlations of perinatal transmission of human immunodeficiency virus type 1 with maternal viremia and lymphocyte phenotypes. J Pediatr
24. Clark SJ, Saag MS, Decker WD, et al. High titers of cytopathic virus in plasma of patients with symptomatic primary HIV-1 infection. N Engl J Med
25. Niu MT, Stein DS, Schnittman SM. Primary human immunodeficiency virus Type 1 infection: review of pathogenesis and early treatment intervention in humans and animal retrovirus infections. J Infect Dis
26. Krivine A, Firtion G, Cao L, Francoual C, Henrion R, Lebon P. HIV replication during the first weeks of life. Lancet