Virus discovery in chronic inflammatory demyelinating polyneuropathy

The events triggering and/or sustaining the auto-immune response underlying chronic inflammatory demyelinating polyneuropathy (CIDP) are unknown. Similar to Guillain-Barré syndrome (GBS), a viral infection might play a role in CIDP. In this study, an virus detection method (VIDISCA-next generation sequencing) capable of detecting known and unknown viruses, was used to analyze the virome in serum of 47 CIDP patients at different time points of the disease and, when available, in cerebrospinal fluid (CSF) samples (N: 17). Serum samples of GBS patients (N:24) and healthy controls (N:114) were used for comparisons. In 5/47 (10.6%; 95% CI: 4-23) CIDP samples, 10/24 (42%; 95% CI: 22-63) GBS samples and 32/114 (28.1%; 95% CI: 20-37) healthy controls samples, anelloviruses were detected, generally regarded as a non-pathogenic species. Parvovirus B19 and GB virus C were found in two CIDP samples (4%). Parvovirus B19, HIV-1 and GB virus C were found in three GBS samples (13%). In 2/17 CIDP CSF samples, an anellovirus and polyomavirus were detected, probably due to contamination during lumbar puncture. No sequences of other viruses were detected in serum or CSF. A (persistent) viral infection sustaining the auto-immune response in CIDP seems therefore unlikely.

Are low and high utilization related to the way GPs manage their practices? An observational study

Background

General practice care plays a key role in keeping healthcare effective and cost-efficient. However, variation in the utilization rates of practices may reveal variation in practice performance. Our research goal is to investigate whether the socio-demographic profile of the patients’ area of residence and practice organization characteristics influence the low or high utilization of general practice care.

Methods

Data on the utilization of general practice care were derived from the electronic health records of 232 general practices participating in the NIVEL Primary Care Database for the year 2013. Census data for the year 2013 were matched with the postal code of the patients. A small area estimation (SAE) technique was used to calculate the estimated utilization rate for general practice care per practice based on the socio-demographic profile of the patients’ area of residence. Subsequently, the actual utilization rates were compared to the estimated rates per practice. Linear regression analysis was used to link the differences between the actual and estimated utilization rates to practice organization characteristics.

Results

The socio-demographic profile of the patients’ area of residence accounted for 25.7% of the estimated utilization rates per practice. Practice organization characteristics accounted for 19.3% of the difference between the actual utilization rates and the estimated rates. Practices had higher utilization rates than estimated when a practice was a dual practice, when it employed female GPs, when it employed other healthcare providers and/or when it offered more services related to a disease management programme.

Conclusion

We found that utilization rates of general practice care can be partially explained by the socio-demographic profile of the patients’ area of residence, but also by practice organization characteristics. Insight into these factors provides both GPs and the other stakeholders involved in the organization of general practice care with information to help reflect on the utilization of care.

Changes in microbiota during experimental human Rhinovirus infection

Background

Human Rhinovirus (HRV) is responsible for the majority of common colds and is frequently accompanied by secondary bacterial infections through poorly understood mechanisms. We investigated the effects of experimental human HRV serotype 16 infection on the upper respiratory tract microbiota.

Methods

Six healthy volunteers were infected with HRV16. We performed 16S ribosomal RNA-targeted pyrosequencing on throat swabs taken prior, during and after infection. We compared overall community diversity, phylogenetic structure of the ecosystem and relative abundances of the different bacteria between time points.

Results

During acute infection strong trends towards increases in the relative abundances of Haemophilus parainfluenzae and Neisseria subflava were observed, as well as a weaker trend towards increases of Staphylococcus aureus. No major differences were observed between day-1 and day 60, whereas differences between subjects were very high.

Conclusions

HRV16 infection is associated with the increase of three genera known to be associated with secondary infections following HRV infections. The observed changes of upper respiratory tract microbiota could help explain why HRV infection predisposes to bacterial otitis media, sinusitis and pneumonia.

Electronic supplementary material

The online version of this article (doi:10.1186/s12879-015-1081-y) contains supplementary material, which is available to authorized users.

Specific serology for emerging human coronaviruses by protein microarray

We present a serological assay for the specific detection of IgM and IgG antibodies against the emerging human coronavirus hCoV-EMC and the SARS-CoV based on protein microarray technology. The assay uses the S1 receptor-binding subunit of the spike protein of hCoV-EMC and SARS-CoV as antigens. The assay has been validated extensively using putative cross-reacting sera of patient cohorts exposed to the four common hCoVs and sera from convalescent patients infected with hCoV-EMC or SARS-CoV.

Clinical manifestations of human coronavirus NL63 infection in children in Taiwan

Human coronavirus NL63 (HCoV-NL63) is a global respiratory tract pathogen; however, the epidemiology of this virus in subtropical area is not well known. To evaluate the epidemics and disease spectrum of HCoV-NL63 infection in children in Taiwan, we prospectively screened children admitted to the hospital with respiratory tract infection from May 2004 to April 2005. Every enrolled child had a nasopharyngeal aspirate (NPA) sample taken. Quantitative RT-PCR was used to detect 1b gene of HCoV-NL63. A total of 539 NPAs were collected. Seven (1.3%) were positive for HCoV-NL63. All cases were boys younger than 3 years of age and most cases occurred in autumn. Co-infection with other pathogens was observed in three cases. The most common symptoms/signs of HCoV-NL63 infection were cough, fever, and inspiratory stridor. HCoV-NL63 was the most common pathogen (14.7%) in children with croup and was the cause of three cases of croup in October. The odds ratio of croup in children infected with HCoV-NL63 was 43.4 (95% CI 8.1 approximately 233.1). In conclusion, HCoV-NL63 is an important respiratory tract pathogen as the main cause in children admitted to the hospital in Taiwan.

Antiviral strategies against human coronaviruses

Since the mid 60's the human coronaviruses (HCoV), represented by HCoV-OC43 and HCoV-229E, were generally considered relatively harmless viruses. This status changed dramatically with the emergence of SARS-CoV in 2002/2003. The SARS-CoV pandemic took 774 lives around the globe and infected more than 8000 people in 29 countries. SARS-CoV is believed to be of zoonotic origin, transmitted from its natural reservoir in bats through several animal species (e.g., civet cats, raccoon dogs sold for human consumption in markets in southern China). The epidemic was halted in 2003 by a highly effective global public health response, and SARS-CoV is currently not circulating in humans. The outbreak of SARS-CoV and the danger of its re-introduction into the human population, as well as the danger of the emergence of other zoonotic coronaviral infections triggered an intense survey for an efficient treatment that resulted in the evaluation of several anticoronaviral compounds. HCoV-NL63 and HCoV-HKU1 were identified shortly after the SARS-CoV outbreak. The 4 human coronaviruses HCoV-229E, HCoV-OC43, HCoV-NL63 and HCoV-HKU1 cause mild respiratory illnesses when compared to SARS, but these infections are involved in 10 - 20 % of hospitalizations of young children and immunocompromised adults with respiratory tract illness. Therefore, there is an urgent need for a successful therapy to prevent disease induction or a vaccine to prevent new infections. This review summarizes the current status of anticoronaviral strategies.

Detection of human coronavirus NL63, human metapneumovirus and respiratory syncytial virus in children with respiratory tract infections in south-west Sweden

Two recently detected viruses, human metapneumovirus (hMPV) and coronavirus NL63 (HCoV‐NL63), have been associated with acute respiratory tract infections, particularly in young children. This study investigated the frequency of hMPV and HCoV‐NL63 infections in Swedish children by screening 221 nasopharyngeal aspirates, collected between November 2003 and May 2005, from 212 children attending the paediatric department of a county hospital in Sweden or submitted from local general practitioners. The samples were originally submitted to be tested for respiratory syncytial virus (RSV), and were examined retrospectively for hMPV and HCoV‐NL63 by RT‐PCR. Of the 212 patients, 101 were positive for RSV (48%), 22 (10%) were positive for hMPV, and 12 (6%) were positive for HCoV‐NL63. The frequency of HCoV‐NL63 infection increased from 1% in 2003–2004 to 10% in 2004–2005. Sequence analysis of parts of the coronavirus genomes showed considerable similarity to the HCoV‐NL63 prototype sequence. The study demonstrated that HCoV‐NL63 and hMPV occur in south‐west Sweden with essentially the same frequency, seasonal distribution and clinical characteristics as have been reported in other countries.

Carrier rate of zidovudine-resistant HIV-1: the impact of failing therapy on transmission of resistant strains

OBJECTIVE

Because maintenance of treatment success in HIV-1 infection requires viruses to remain therapy sensitive in drug-naive seropositive persons, we looked at the primary infections caused by drug-resistant HIV-1 over time. Furthermore, to study the coverage rate of therapy and therapy failure in relation to the transmission of resistant viruses a mathematical model was developed.

DESIGN

The reverse transcriptase and protease genes of viruses were analysed in newly infected people in the period 1990-1998 in the Amsterdam Cohort Study on HIV infection and AIDS in homosexual men.

METHODS

The mathematical model was based on the coverage of drug regimens selecting zidovudine (ZDV) resistance, the lag time in which resistance is gained or lost, the death rate of people infected with resistant virus, and the replacement of resistance-selecting regimens by more potent treatments that substantially reduce viral load and mortality.

RESULTS

Of 43 individuals with a primary HIV-infection, three (7%) harboured ZDV-resistant viruses. The first of the ZDV-resistant strains was transmitted in 1995, the last two in 1996. The build-up of ZDV resistance was described by the mathematical model indicating that the equilibrium level of resistance due to treatment depends only on the treatment rate and the outflow rate of patients with resistance virus.

CONCLUSIONS

Our model indicates that the frequency of viral resistance in a population is determined largely by the number of individuals on insufficient or failing therapy and is influenced only modestly by secondary transmission of ZDV-resistant strains.

Slower decline of plasma HIV-1 RNA following highly suppressive antiretroviral therapy in primary compared with chronic infection

OBJECTIVES

To study the effect of highly suppressive antiretroviral therapy on the slopes of HIV-1 RNA decline in primary compared with chronic HIV-1 infection.

METHODS

Slopes of HIV-1 RNA decline in plasma were compared before and after the start of highly suppressive antiretroviral therapy from five acutely infected patients who started treatment 2 to 5 weeks following the onset of clinical symptoms. Slopes of decline after the initiation of therapy were also compared with those found in 12 chronically infected individuals on the same therapy. Numbers and percentages of activated CD4 and CD8 T cells at baseline were compared as well.

RESULTS

The pre-treatment slopes of HIV-1 RNA decline in the acutely infected individuals increased significantly (P = 0.0001) after the start of anti-retroviral therapy. However, these post-treatment slopes were lower than those found in the chronically infected individuals (P= 0.012). Slopes were inversely correlated (P= 0.012) with baseline HIV-1 RNA. Although the number of CD38+HLA-DR+ CD4 cells was higher in primary infection (P= 0.02), the percentage did not differ between primary and chronic infection.

CONCLUSIONS

These findings indicate that antiretroviral therapy contributes significantly to the clearance of HIV-1 during primary infection. Based on the mathematical model the less steep RNA slope following the start of treatment in primary infection can be predicted to be the result of lower clearance of productively infected cells and higher burst size per cell per unit time. This may indicate a growing immune response to HIV-1 in this very early stage of infection.

Human immunodeficiency virus type 1 in faeces and serum: evidence against independently evolving subpopulations

It is not known whether independent tissue-specific evolution accounts for the differences between human immunodeficiency virus type 1 (HIV-1) subpopulations in intestinal tissue and blood. To study this, sequential serum samples from three persons were analysed for the presence of HIV-1 V3 genotypes which were detected exclusively in faeces at a specific time-point. For two persons the faeces genotype was found in serum samples collected before the time of faeces collection: 7 months for one person and 32 months for the other person. In the third person, serum collected 1 month after faeces collection contained the faeces genotype in abundance. These data indicate that a difference between intestinal tissue and blood HIV-1 subpopulations is not the result of complete compartmentalization and independent HIV-1 evolution in intestinal tissue, but that it reflects an unequal distribution of HIV-1 in different tissues.

Genetic differences between human immunodeficiency virus type 1 subpopulations in faeces and serum

To study human immunodeficiency virus type 1 (HIV-1) compartmentalization between intestine and blood, paired faecal and serum samples were collected from 204 HIV-1-infected persons. Direct sequencing of the gp120 V3 region obtained from 33 persons showed that faecal and serum sequences could be nearly homologous (0.3% different) or very dissimilar (11.3% different). Individual clones were obtained and sequenced from the faecal and serum samples of 13 persons. In 6 persons the HIV-1 subpopulations in faeces and serum were similar, whereas in 7 persons, distribution of V3 genotypes showed a marked difference. Genetic characterization of the HIV-1 subpopulations showed less heterogeneity in faecal subpopulations than in serum subpopulations in 5 of the 7 subjects. Furthermore, faecal and serum subpopulations differed predominantly by nonsynonymous nucleotide substitutions (in 6 of 7 persons). Comparison of the HIV-1 subpopulations in faeces and serum of these 7 persons, using resampling techniques, revealed a significant difference between faecal and serum subpopulations at an N-linked glycosylation site, C-terminal of the V3 loop (amino acids 331-333). Sequences from faecal subpopulations of all 7 persons contained a glycosylation site at amino acid position 331-333. Four of these 7 harboured serum variants lacking a glycosylation site at this position. The faecal subpopulations in these 4 persons showed limited nonsynonymous substitutions compared to synonymous substitutions, indicating that purifying selection is operational on these subpopulations.

Prevalence of Campylobacter-associated diarrhea among patients infected with human immunodeficiency virus

We performed a cross-sectional study at an outpatient AIDS clinic to assess the prevalence of Campylobacter species in stool specimens from 201 consecutive patients infected with human immunodeficiency virus (HIV). We characterized campylobacters phenotypically and genetically by using primers for the group of common species (i.e., C. jejuni, C. coli, C. lari, and C. upsaliensis) and for most individual uncommon species. We performed cultures with use of a membrane filter technique on nonselective blood agar and found that Campylobacter species were the most frequent enteropathogenic bacteria: the organisms were recovered from 7 (16%) of 43 patients with diarrhea and 5 (3%) of 158 patients without diarrhea (P = .001). We isolated only one campylobacter with use of conventional culture techniques on selective media. Phenotypic characterization of 10 campylobacter strains resulted in the misidentification of four isolates. C. upsaliensis was the most frequently isolated species, followed by C. jejuni and C. coli. Two strains could not be identified with the available primers. Two of 12 Campylobacter strains were resistant to erythromycin, and two were resistant to ciprofloxacin. We conclude that Campylobacter species other than C. jejuni can frequently be detected in the stools of HIV-infected patients and that these organisms could be associated with diarrhea.

Human immunodeficiency virus type 1 RNA populations in faeces with higher homology to intestinal populations than to blood populations

To determine whether human immunodeficiency virus type 1 (HIV-1) in faeces is representative of the HIV-1 population in intestinal tissue, we studied HIV-1 V3 variation in faeces, intestinal biopsies and serum from two individuals. Phylogenic analysis of HIV-1 V3-coding RNA in faeces from one individual showed three distinct genotypes. Viruses belonging to all three genotypes were also present in sigmoidal tissue and in serum. Jejunal tissue contained two of these three genotypes. Analysis of the V3-coding RNA in faeces of the other individual showed five distinct genotypes. One of these genotypes was present in all specimens from this individual. Besides this shared genotype, jejunal tissue and serum contained sequences belonging to one other genotype. In addition, one of the other three V3 variants was detected in sigmoidal tissue. For both persons the shared HIV-1 RNA genotypes in faeces and serum displayed a distinctly different frequency distribution. In one individual, the genotype which was detected in a majority of the clones in faeces (59%) and as a minority in serum (11%), was the most abundant genotype in jejunal and sigmoidal tissue (61% and 80%, respectively). For the other individual the genotype that was present in faeces in a significant number of clones (43%) was detected in serum as a minority (8%), whereas this genotype composed 47% of the clones isolated from jejunal tissue. Taken together these data suggest that faeces contain HIV-1 sequences that are derived from local HIV-1 replication in intestinal tissue.

Syncytium-inducing (SI) phenotype suppression at seroconversion after intramuscular inoculation of a non-syncytium-inducing/SI phenotypically mixed human immunodeficiency virus population

Two distinct biological phenotypes of human immunodeficiency virus (HIV) have been described: the non-syncytium-inducing (NSI) phenotype, best characterized by the inability to infect MT-2 cells, and the syncytium-inducing (SI) phenotype, with the ability to infect MT-2 cells. The earliest virus population observed following HIV transmission is generally of the NSI phenotype, even after exposure to inocula of mixed NSI/SI phenotype. In this study, the issue of intrapatient selection of virus phenotype following transmission was addressed by studying two cases of accidental transmission. A comparison of the sequences of the V1-V2 and the V3 coding regions of the envelope gene and the p17 region of the gag gene showed that the donor-recipient pairs were tightly clustered in all gene segments, but away from local and published transmission controls. The intrasample variation of the p17 sequence was greater in the recipients and smaller in the donors than that of the V3 region sequence, indicating selection of V3 at transmission. In these transmission cases, the effects of an intravenous inoculation of a small quantity of blood containing predominantly SI V3 sequences (6 of 8 clonal sequences) were compared with those of an intramuscular inoculation of a large quantity of blood containing predominantly NSI viruses (14 of 16 clonal sequences). Both SI and NSI V3 regions were demonstrated to be phenotypic expressions of genetically related viral strains. The inoculation of the predominantly SI virus population resulted in the persistence of an SI virus population in the recipient and a rapid CD4+ T-cell decline. The inoculation of the predominantly NSI population resulted in a selective amplification of SI viruses before seroconversion, followed by a suppression of SI viruses at seroconversion and a rapid decline of CD4+ T-cell numbers. These data suggest that the suppression of SI viruses can be accomplished following the development of HIV-specific immunity and that the ability to suppress SI viruses does not prevent the development of immunodeficiency.

Isolation of human immunodeficiency virus type 1 (HIV-1) RNA from feces by a simple method and difference between HIV-1 subpopulations in feces and serum

A simple method for the isolation and subsequent detection of human immunodeficiency virus type 1 (HIV-1) RNA from feces is described. Viral RNA was isolated by the method developed by Boom et al. (R. Boom, C.J.A. Sol, M.M.M. Salimans, C.L. Jansen, P.M.E. Wertheim-van Dillen, and J. van der Noordaa, J. Clin. Microbiol. 28:495-503, 1990), which was adapted for feces. HIV-1 RNA was detected by reverse transcription (RT) followed by a nested PCR encompassing the V3 region. Reconstruction experiments revealed that the efficiencies of the extraction technique and the subsequent RT-PCR were not considerably affected by the varied composition of feces. The method was applied on fecal specimens from 18 HIV-1-infected individuals, among which were samples that had been stored for 9 years. It appeared that HIV-1 RNA was detectable in the feces of 12 persons (67%). Viral RNA was present in the feces of persons who fulfilled the criteria for CDC class II and CDC class III HIV infection as well as in patients who were diagnosed with AIDS (CDC class IV). Direct sequencing of amplimers obtained from paired fecal and serum specimens showed that differences in sequence heterogeneity existed. In one patient a remarkable difference in the HIV-1 sequences between isolates from feces and serum was observed. In conclusion, HIV-1 RNA is frequently present in the feces of HIV-1-infected individuals, and in some cases the HIV-1 subpopulation in feces differs from the HIV-1 subpopulation in serum.

Antibody responses to HIV-1 envelope and gag epitopes in HIV-1 seroconverters with rapid versus slow disease progression

We studied the relationship between the rate of disease progression after HIV-1 seroconversion and the level of IgG antibody response to HIV-1 envelope and core epitopes. This was done by comparing a group of fast-progressing individuals and a group of slow-progressing individuals for serum IgG titers to peptides from the gp120-V3 neutralization domain, to a peptide from the immunodominant gp41 epitope (residues 590 to 607), and to recombinant gp120 and p24. The two groups displayed a large overlap in titers to the envelope epitopes, which precluded their differentiation at most time points after seroconversion. Low responsiveness to envelope antigens was not only found in a few fast-progressors but also in one individual who remained asymptomatic for at least 92 months after seroconversion. The only significant differences between the groups were found in the first months after seroconversion when the responses to the V3 domain and the gp41 epitope were more vigorous in the group of fast-progressors. Furthermore, on evaluating ratios of anti-V3 antibody titers to anti-gp120 antibody titers we found no indication that fast disease progression was associated with a restriction in antibody response to the V3 epitope. We did confirm the finding that fast disease progression is associated with low levels of p24-directed antibodies, both early after seroconversion and at later stages. These data demonstrate that levels of IgG antibodies to envelope epitopes are poor predictors of rapid disease progression and suggest that the role of V3-directed neutralizing antibodies in preventing subversion of the immune system is not decisive in natural HIV-1 infection.

Frequent and early HIV-1MN neutralizing capacity in sera from Dutch HIV-1 seroconverters is related to antibody reactivity to peptides from the gp120 V3 domain

The temporal development of HIV-1 neutralizing activity and antibodies to the gp120-V3 neutralization domain were studied in sera from 20 Dutch HIV-1-infected individuals followed from seroconversion on. Serum neutralizing capacity was assessed with three T cell line-tropic isolates: HIV-1MN, HIV-1HXB2, and the patient isolate HIV-1(320). Neutralizing activity to HIV-1MN developed in 18 individuals (90%) within 0 to 10 months after seroconversion. Parallel evolution of IgG reactivity to V3 peptides of United States/European type variants, and the capability of such peptides to completely inhibit HIV-1MN neutralization in four of five tested sera (taken 1-2 years after seroconversion), indicate that a large proportion of HIV-1MN neutralizing antibodies is directed to V3. The early appearance and high frequency of HIV-1MN neutralizing activity in the Dutch study group indicate the close relationship of HIV-1MN to HIV-1 variants circulating in the Netherlands. Neutralizing activity to HIV-1HXB2 (in 15 of 20 individuals) developed several months after that to HIV-1MN in all individuals (average, 10 months after seroconversion) and was not seen in the absence of HIV-1MN neutralizing activity. Neutralizing activity to the Dutch isolate HIV-1(320) (found in 11 of 18 tested individuals) emerged simultaneously with that to HIV-1MN in 4 individuals but appeared later in 7. In most individuals, HIV-1HXB2 neutralization was not accompanied by reactivity to a V3 peptide from this strain, indicating that the extension of neutralizing activity to more divergent strains, which takes place at later stages, must be attributed to non-V3-directed antibodies.

Immunodominance and antigenic variation of the principal neutralization domain of HIV-1

The principal neutralization domain (PND) of HIV-1 is located in the third variable region (V3) of the envelope glycoprotein gp 120. Cross-reactivity of experimental and natural sera with recombinant proteins containing the V3 region of four HIV-1 variants showed that a group of viruses (among which HIV-1 MN) had antigenically similar V3 regions. The V3 regions of HIV-1 IIIB and HIV-1 RF were antigenically distinct. Antibodies raised to V3 domains of two isolates from the "MN group" neutralized HIV-1 MN (and not HIV-1 IIIB), thus confirming the antigenic similarity of V3 of these isolates to that of HIV-1 MN. Human antibodies to the V3 region were shown to be mainly directed to the central area in V3, where the neutralization domain is. In addition, antibody reactivities in sera of 397 Dutch and 39 Tanzanian HIV-1-infected individuals show that the V3 neutralization domain is highly antigenic, and that viruses from the MN group predominate in The Netherlands and to a lesser extent in Tanzania. Thus, if the protective value of antibodies to the PND can be established, then PND (poly)peptides derived from the MN group may be important components of a subunit vaccine against HIV-1.

Human antibody response to a strain-specific HIV-1 gp120 epitope associated with cell fusion inhibition

PEPSCAN analysis, performed using 536 overlapping nonapeptides derived from the HTLV-III B nucleotide sequence of the region encoding the external envelope protein of 120 kDa (gp120), identified in the V3 region of gp120 a major binding site for antibodies of HIV-1-infected humans. The minimal amino acid sequence of this antibody binding site was demonstrated by multiple length scanning to be five to eight amino acids in length: (G)PGRAF(VT), i.e. amino acids 312-319. A peptide (Neu 21) containing this binding site for human antibodies (KSIRIQRGPGRAFVTIG) was synthesized and shown to induce HTLV-III B cell fusion-inhibiting antibodies in rabbits and mice. Antibodies binding to this HTLV-III B/LAV-1-specific peptide were shown to be primarily of the IgG 1 subclass, appeared within 6 months after HIV-1 antibody seroconversion in six out of 14 men studied, and persisted throughout the follow-up period of 10-24 months. The other eight seroconverting men did not develop antibodies to Neu 21 during the observation period. The appearance of antibodies to Neu 21 paralleled the capacity of the serum to inhibit HTLV-III B in cell fusion. HIV-1-infected men with Kaposi's sarcoma exhibited a similar frequency of antibodies to the synthetic peptide Neu 21 (14 out of 39, 36%) as asymptomatic HIV-1-infected men (112 out of 319, 35%). Adults with Pneumocystis carinii pneumonia had a significantly lower frequency (11 out of 78, 14%) of antibodies to Neu 21. Similarly, a low prevalence of antibodies to Neu 21 (8 out of 43, 19%) was observed among symptomatic HIV-1-infected children.