Pneumococcal vaccination in adult people living with HIV on suppressive antiretroviral therapy: a case-control study

There is little information on the effectiveness of the pneumococcal vaccines (PVs), especially for the pneumococcal conjugate vaccines (PCVs), in HIV-infected patients in the modern antiretroviral era. This is a case–control study where cases were people living with HIV (PLWH) with confirmed pneumococcal infection (CPI) and controls were PLWH without CPI matched with cases by gender and year of HIV diagnosis. The selection process was blinded to the study factor (vaccination). Sample size estimation yielded 61 cases and 183 controls. We analyzed the effect of PV on CPI using Cox proportional-hazards regression model with time-dependent covariates. We included 256 subjects: 64 cases, and 192 controls. PVs had been administered to 115 (45%) patients. Only the modified Charlson Comorbidity Index (HR 1.16, 95%CI 1.06–1.27, P = 0.001) and the CD4 nadir (HR 0.99, 95% CI 0.98–0.99, P = 0.001) were independently associated with CPI. Receipt of PV was not associated with CPI after adjusting in the multivariate model with time protection as a dependent covariate (HR 0.65, 95% CI 0.35–1.32 P = 0.250). We also investigated the influence of different immunization schedules. In an adjusted model, we found no evidence of protection against CPI, including double immunization schedules (HR 0.42 95%CI 0.15–1.19 P = 0.102). In this case–control study, we could not show an association between pneumococcal vaccination and confirmed pneumococcal infection, although a protective effect of particular schedules of immunization cannot be ruled out.

Genotypic tropism testing of proviral DNA to guide maraviroc initiation in aviraemic subjects: 48-week analysis of results from the PROTEST study

OBJECTIVES

Maraviroc (MVC) is a suitable drug for aviraemic subjects on antiretroviral treatment (ART) developing toxicity. Its prescription requires prior tropism testing. It is unknown if proviral DNA genotypic tropism testing is reliable for guiding MVC initiation in aviraemic subjects, so this study was carried out to address this issue.

METHODS

PROTEST was a phase 4, prospective, single-arm clinical trial carried out in 24 HIV care centres in Spain. MVC-naïve HIV-1-infected patients with HIV-1 RNA < 50 copies/mL on stable ART during the previous 6 months who required an ART change because of toxicity and who had R5 HIV, as determined by proviral DNA genotypic tropism testing, initiated MVC with two nucleoside reverse transcriptase inhibitors (NRTIs) and were followed for 48 weeks. Virological failure was defined as two consecutive viral load measurements > 50 copies/mL.

RESULTS

Tropism results were available for 141 of 175 (80.6%) subjects screened: 60% had R5 and 85% of these (n = 74) were finally included in the study. Previous ART included protease inhibitors (PIs) in 62% of subjects, nonnucleoside reverse transcriptase inhibitors (NNRTIs) in 36%, and integrase inhibitors (INIs) in 2%. Main reasons for treatment change were dyslipidaemia (42%), gastrointestinal symptoms (22%) and liver toxicity (15%). MVC was given alongside tenofovir (TDF)/emtricitabine (FTC) (54%) and abacavir (ABC)/lamivudine (3TC) (40%) in most patients. Eighty-four per cent of patients maintained a viral load < 50 copies/mL to week 48, whereas 16% discontinued treatment: two withdrew informed consent, one had an R5 to X4 shift between screening and baseline, one was lost to follow-up, one developed an adverse event (rash), two died from non-study-related causes, and five developed protocol-defined virological failure.

CONCLUSIONS

Initiation of MVC plus two NRTIs in aviraemic subjects based on genotypic tropism testing of proviral HIV-1 DNA is associated with low rates of virological failure for up to 1 year.

Incidence and prognosis of immune reconstitution inflammatory syndrome in HIV-associated progressive multifocal leucoencephalopathy

BACKGROUND AND PURPOSE

Progressive multifocal leucoencephalopathy-associated immune reconstitution inflammatory syndrome (PML-IRIS) is the paradoxical worsening or unmasking of preexisting infection with JC virus attributable to a rapid recovery of the immune system after highly active antiretroviral therapy (HAART) initiation. We investigated the incidence and factors associated with PML-IRIS in HIV-infected patients. We also studied its influence on mortality of PML and the effect of corticosteroid therapy.

METHODS

Single-center retrospective analysis of HIV-infected patients diagnosed with PML from 1996 to 2012 who received HAART.

RESULTS

Among 59 PML patients treated with HAART, 18 (30.51%) developed PML-IRIS (five delayed PML-IRIS, 13 simultaneous PML-IRIS). Patients who developed IRIS had lower CD4 counts prior to treatment (102 vs. 68.5, P < 0.05) and experienced a greater decline in HIV-RNA levels in response to HAART (2.5log vs. 2.95log, P < 0.05). Gadolinium enhancement on MRI was observed in 31.25% of PML-IRIS cases versus 2.56% of PML non-IRIS (P < 0.01). Survival rates were higher in patients with PML-IRIS compared to those with PML non-IRIS. Eight patients received corticosteroids, five of which had a good outcome. Patients who died were severely ill when treatment was initiated whereas patients who survived were treated before major neurological deterioration occurred.

CONCLUSIONS

Nearly one-third of HIV-infected patients with PML develop IRIS after initiating HAART. Patients severely immunocompromised who experience a rapid virological response to HAART have a higher risk for PML-IRIS. There was a trend for lower mortality in patients with IRIS. Early treatment with corticosteroids might be useful.

Visceral leishmaniasis as an independent cause of high immune activation, T-cell senescence, and lack of immune recovery in virologically suppressed HIV-1-coinfected patients

OBJECTIVES

Different immune alterations have been described in HIV-infected patients with visceral leishmaniasis (VL). We aimed to identify the immunological factors involved in the lack of immunological recovery and VL relapses in HIV-infected patients with VL, by comparison with other HIV-infected patients.

METHODS

We carried out a cross-sectional study of 55 patients receiving suppressive combination antiretroviral therapy (cART) for at least 1 year: nine with previous relapsing VL, 20 with an immunodiscordant response (IDR) to cART (CD4 count < 200 cells/μL) and no previous VL, and 26 with a concordant response (CR) to cART (CD4 count > 350 cells/μL) without VL. Immunosenescence was investigated by analysing CD57(+) CD28(-) levels, immune activation by analysing CD38(+) HLA-DR(+) levels, inflammation by analysing interleukin (IL)-6 levels, and microbial translocation by analysing lipopolysaccharide (LPS) and soluble CD14 (sCD14) levels.

RESULTS

In VL patients, the median time since VL diagnosis was 42 months, and all patients had had at least one relapse despite suppressive cART for a median time of 43 months. Patients with previously diagnosed VL had a higher CD8 T-cell activation level (P < 0.001) than those with IDR. Also, levels of IL-6, LPS and especially sCD14, associated with bacterial translocation and additional monocyte activation, were significantly increased in patients with previous VL compared with patients with IDR (P = 0.048, P = 0.049 and P < 0.001, respectively). In addition, patients with previous VL had higher levels of CD8 T-cell senescence. Notably, the levels of immune activation and inflammation in patients with previous VL were not related to the time of VL diagnosis, the number of VL relapses, or hepatitis C virus (HCV) coinfection.

CONCLUSIONS

Our data demonstrate that VL patients had an even worse immunological status than patients with IDR, which was probably associated with increased microbial translocation and additional monocyte/macrophage activation. These data explain the observed lack of immunological recovery and the occurrence of VL relapses in HIV-infected patients with previous VL.

High levels of CD4⁺ CTLA-4⁺ Treg cells and CCR5 density in HIV-1-infected patients with visceral leishmaniasis

Visceral leishmaniasis (VL) in HIV-1-infected patients has been associated with poor immunological recovery and frequent disease relapses. The aim of this study was to analyse the role of T cell populations, Treg cells and CCR5 density in patients with VL compared to HIV-1-infected patients without leishmaniasis. A cross-sectional study of nine Leishmania-HIV-1-coinfected (LH) patients with VL receiving suppressive cART for at least 1 year were compared to 16 HIV-1-infected patients with non-immunological response (NIR, CD4 count below 250 cells/mm(3)) and 26 HIV-1-infected patients with immunological response (IR, CD4 count above 500 cells/mm(3)) without leishmaniasis. LH patients had a deep depletion of naïve T cells (p = 0.002), despite similar levels of effector T cells compared to NIR patients. CD4 Treg cells were similar compared to NIR patients, but higher compared to IR patients (p < 0.001). Interestingly, CD4 Treg CTLA-4(+) cells were higher in LH patients compared to either NIR or IR patients (p = 0.022 and p < 0.001, respectively), and the CD4 Treg/TEM ratio was similar to NIR patients, but higher compared to IR patients (p = 0.017). CCR5(+) T cell levels were higher compared to IR patients (p < 0.001), while CCR5 density on T cells were higher compared to both NIR and IR patients (p < 0.005 in both cases). Higher levels of CD4(+) CTLA-4(+) Treg cells and CCR5 density on CD8(+) T cells are strongly associated with VL in HIV-1-infected patients. Also, these patients have a poor immunological profile that might explain the persistence and relapse of the pathogen.

Regression of liver fibrosis is progressive after sustained virological response to HCV therapy in patients with hepatitis C and HIV coinfection

There are few data about the long-term histological outcome of HIV-/HCV-coinfected patients after therapy with interferon and ribavirin. We performed an observational study of 216 patients who received therapy against HCV and who had at least three successive transient elastographies (TE) during the follow-up. The primary endpoint was confirmed fibrosis regression, defined as a reduction of at least 1 point in Metavir fibrosis score, confirmed and without worsening in successive TE. At baseline, 23% had fibrosis stage 4 or cirrhosis. Overall, 82 (38%) achieved sustained virological response (SVR), without differences in baseline fibrosis or time of follow-up. Confirmed fibrosis regression was observed in 55% of patients, higher for SVR (71% vs 44%; P < 0.01), and the likelihood of achieving fibrosis regression at 3, 5 and 7 years was 0.17, 0.51 and 0.67, respectively, for SVR patients, in comparison with 0.02, 0.23 and 0.41 for no SVR patients (P < 0.01, log-rank test at any time point). Progressive regression, defined as continuous improvement in successive TE, was observed in 62% of patients with advanced liver fibrosis or cirrhosis who achieved SVR. In a Cox regression model, only SVR (HR, 4.01; 95% CI, 2.33-7.57; P < 0.01) and a younger age (HR, 1.14; 95% CI, 1.05-1.25; P < 0.01; per year) were associated with fibrosis regression. This study confirms that the rate of liver fibrosis regression increases during the follow-up after SVR to interferon therapy in HIV-/HCV-coinfected patients.

Viral kinetics and early prediction of nonresponse to peg-IFN-alpha-2b plus ribavirin in HCV genotypes 1/4 according to HIV serostatus

To evaluate, among 70 hepatitis C virus (HCV)-monoinfected and 36 human immunodeficiency virus (HIV)-coinfected naïve patients with genotypes 1/4 receiving weight-adjusted pegylated interferon-alpha-2b/ribavirin, viral kinetics and the feasibility to predict treatment failure measuring early HCV-RNA decreases. HCV-RNA was assessed at baseline, weeks 4, 12 and 24. Receiver operating characteristic (ROC) curves were calculated to determine the most sensitive cut-off values of viral decrease at week 4 predicting treatment failure. Baseline predictors of failure were evaluated by univariate and multivariate analyses. Despite similar baseline HCV-RNA (5.75 vs 5.72 log(10)IU/ml, P = 0.6), HCV monoinfection led to significantly lower HCV-RNA values at weeks 4 (3.7 vs 4.3 log(10)IU/ml, P = 0.01), 12 (2.3 vs 3.5 log(10)IU/ml, P = 0.01) and 24 (1.4 vs 3.3 log(10)IU/ml, P = 0.001) and a higher rates of viral clearance at weeks 24 (60%vs 36%, P = 0.02), 48 (46%vs 25%, P = 0.03) and 72 (37%vs 17%). The lack of achieving an HCV-RNA decrease of at least 1 log(10) at week 4 was highly predictive of treatment failure for HCV-monoinfected patients (Se 100%, Sp 50%, positive predictive value (PPV) 57%, negative predictive value (NPV) 100%, ROC curve area, 0.86 [95% confidence interval (CI) 0.77-0.95], but not for HCV/HIV-coinfected patients (cut-off, 0 log(10), Se 100%, Sp 27%, PPV 21%, NPV 100%, ROC curve area, 0.71 (95% CI 0.49-0.93). HIV coinfection was independently associated with failure (odds ratio 2.95, 95% CI 1.08-8.04, P = 0.01). Thus the magnitude of HCV-RNA decreases at week 4 correlated with treatment response. Significant differences in viral kinetics and cut-off values predicting nonresponse suggest a slower HCV clearance rate in HIV coinfection, which was independently associated with treatment failure.

[Cytomegalovirus sensitivity to anti-retroviral agents in patients infected with the human immunodeficiency virus with chorioretinitis in Spain]

OBJECTIVES

To determine the prevalence of the in vitro resistance rate of cytomegalovirus (CMV) to gancyclovir (GCV) and foscarnet (FOS) among patients with AIDS and chorioretinitis. Furthermore, in vitro sensitivity results were compared with the clinical response to therapy.

PATIENTS AND METHODS

Thirty-six patients with the diagnosis of CMV retinitis and AIDS were included in the study. Antiviral sensitivity testing was performed to 51 clinical CMV strains from these patients. The in vitro sensitivity was compared with the clinical response to therapy. The resistance criteria were the inhibitory dose50 (ID50) GCV > 5 M and ID50 FOS > 400 M.

RESULTS

None of the CMV tested strains was resistant to GCV or FOS; however, six patients who had relapses of chorioretinitis while on maintenance therapy and isolation during the condition had mean ID50 of strains (n=8) of 1,95 M for GCV (standard deviation [SD] of 0.71) and 115.2 for FOS (SD, 34.7). These patients responded well when drugs were used at induction doses.

CONCLUSION

The in vitro sensitivity testing is not a good predictor of infection control when the drug is used at maintenance doses. This suggests that these strains should be classified in the intermediate resistant category.

Drug resistance in patients experiencing early virological failure under a triple combination including indinavir

OBJECTIVE

To assess the pattern of drug resistance mutations selected in HIV-1-infected patients failing a first line triple combination therapy including indinavir.

PATIENTS AND METHODS

Plasma samples from 87 patients collected at the time of the first virological rebound (> 50 HIV-RNA copies/ml) were examined for the presence of drug-resistant genotypes.

RESULTS

The mean level of plasma viraemia at rebound was 7824 HIV-1 RNA copies/ml in 73 subjects with good compliance, whereas it was 359,460 HIV-1 RNA copies/ml in 14 patients who admitted to poor adherence. Genetic sequence analysis yielded results for 51 (70%) of the patients having good adherence. More than half of them (26/51, 51%) carried primary mutations associated with resistance to nucleoside analogues. In contrast, primary protease inhibitor resistance mutations were recognized less frequently (14/51, 27%; P < 0.05). Moreover, in 23 (45%) patients there was no evidence of drug-resistant viruses at all. The most frequent drug-resistant genotypes in the reverse transcriptase gene were at codons 184 (n = 19), 215 (n = 14) and 41 (n = 8), whereas for the protease they were at codons 46 (n = 10), 82 (n = 9) and 90 (n = 7). No resistance genotypes were found among non-compliant patients.

CONCLUSION

The overall rate of drug-resistant HIV genotypes was 38% (28/73) in patients with good adherence and who were experiencing a first virological failure under a triple combination regimen including indinavir; resistance to nucleoside analogues was more frequent than resistance to indinavir. Therefore, treatment intensification in those patients without resistance, or a selective substitution of nucleosides in those with resistance limited to these compounds, might be justified.

Phenotypic cross-resistance to nelfinavir: the role of prior antiretroviral therapy and the number of mutations in the protease gene

Cross-resistance to nelfinavir (NFV) is observed in patients failing protease inhibitor (PI)-containing therapies. We performed a study with 111 patients who started an NFV-based salvage regimen after failing PI-based therapy to evaluate genotypic changes and to identify factors associated with resistance to NFV. Genotypic and phenotypic resistance data at entry (111 and 51 samples) and after NFV failure (74 and 31 samples) were available. Median CD4(+) cell count was 208 x 10(6)/liter, HIV RNA level was 4.6 log(10) copies/ml, and median number of mutations in the protease was 9. At baseline, 51 and 14% of viral isolates showed high or intermediate phenotypic resistance to NFV. Phenotypic data correlated with virological outcome, reaching undetectability at the third month in 40, 14, and 0% of those patients with susceptible, intermediate, or resistant viral isolates, respectively. Phenotypic resistance to NFV was associated with the presence of the L90M mutation: 46% for resistant vs. 6% in susceptible strains. The number of mutations in the protease correlated with the fold-increase in the IC(50)-NFV. The D30N mutation was detected in only 1 of 74 patients who failed. In a logistic regression analysis, the number of mutations in the protease was associated with NFV cross-resistance (RR, 2.09 per each additional mutation; 95% CI 1.23-3.55; p < 0.01). In conclusion, phenotypic cross-resistance to NFV for PI-experienced patients can be predicted by the number of mutations in the protease. The L90M mutation is significantly associated with the subsequent failure of NFV-containing regimens. The presence of the D30N mutation was rare and not useful in identifying NFV-resistant isolates.

Phenotypic testing predicts virological response in successive protease inhibitor-based regimens

OBJECTIVE

To evaluate the importance of the number of active drugs, as determined by phenotypic resistance testing, in achieving virological response in successive salvage regimens.

DESIGN

Phenotypic study of 57 plasma samples corresponding to 24 patients who had sequentially received three protease inhibitor-containing regimens. Phenotypic susceptibility to a drug (active drug) was defined as less than a four-fold-increase in the IC50 in comparison with the wild type.

MAIN OUTCOME MEASURE

Virological response according to the number of active drugs (three versus two or fewer), HIV load, length of antiretroviral exposure, and line of protease inhibitor-based therapy (first, second and third regimen).

RESULTS

Before the first protease inhibitor-based therapy, the median time on antiretroviral treatment was 42 months, and before the second and third protease inhibitor-salvage regimens it was 10 and 8 months, respectively. The number of patients receiving three active drugs simultaneously was 24, 35 and 31% in each line of therapy. At week 12, a close correlation was found between the presence of three active drugs in the antiretroviral regimen and the rate of virological response, in comparison with those patients receiving two or fewer active drugs [76 versus 45%, relative risk (RR), 1.7; 95% confidence interval (CI) 1.1-2.6; P = 0.028]. In a multivariate analysis, the use of two or fewer active drugs was an independent predictor of lack of response, regardless of HIV load, length of previous antiretroviral exposure and line of salvage therapy (RR, 4.5; 95%CI, 1.1-18.3; P = 0.03). Of note, a higher rate of response was observed in patients receiving the first protease inhibitor-containing regimen in comparison with those in subsequent protease inhibitor-based salvage regimens (83 versus 50 versus 28%, P < 0.01), even when only those patients receiving three active drugs were included (100 versus 71 versus 60%).

CONCLUSIONS

This data confirm the usefulness of phenotypic testing in guiding antiretroviral therapy in heavily pretreated patients. The number of active drugs and the line of salvage therapy are independent predictors of virological response, regardless of HIV load and the length of antiretroviral exposure.

Patients failing saquinavir therapy require an early change to indinavir before HIV-1 viral load reaches high levels

Sequential use of antiretroviral therapy with protease inhibitors (PI) is frequently prescribed owing to failure or intolerance of the first selected agent. Controversial data exist about the virological and immunological outcome of patients in whom a change to a second PI regimen is needed. A prospective study of 113 HIV-positive patients (male, 84%; mean age 36 years; previous AIDS-defining event, 35%; previous antiretroviral therapy with nucleoside analogues, 94%) who started a saquinavir-containing regimen between March 1996 and March 1997 and had to change to indinavir (n = 79) owing to intolerance, failure or medical criteria was performed. At the time of the switch, median CD4 cell count was 221 cells/mm3 and the HIV RNA level was 3.98 log10 copies/ml. The rate of viral suppression (HIV RNA levels below 200 copies/ml) was 40% at 3 months and 28% at month 6 after indinavir was instituted. In a logistic regression analysis, only the baseline viral load [relative risk (RR), 2.85; 95% confidence interval (CI), 1.31-6.05; P = 0.007] was statistically associated with the lack of viral suppression on indinavir. Although there are not sufficient data about the best therapeutic option if a change in PI-containing regimens therapy is considered, this study supports the recommendation of an early change of the PI-containing regimens, before the HIV-1 viral load reaches high levels.

Acute polyradiculopathies in HIV-infected patients

We studied 17 consecutive cases of acute polyradiculopathy (PR) diagnosed in HIV-infected patients to investigate the possible causes of this syndrome in our milieu. Sixteen patients presented with lumbosacral PR and one patient had predominantly cervical PR. Electrophysiological study showed a predominantly motor axonal neuropathy in all patients examined. Six patients had a laboratory-confirmed aetiology for the PR: cytomegalovirus (CMV) was isolated from cerebrospinal fluid (CSF) in three cases, meningeal lymphomatosis was diagnosed by CSF cytology in two cases, and one patient had cryptococcal meningitis. Another patient was thought to have acute axonal polyradiculoneuritis associated with HIV infection. CMV and Mycobacterium tuberculosis were the probable agents in four and three patients, respectively. Finally, in three patients a cause could not be foscarnet were effective in the treatment of definite or probable CMV PR. The present study confirms that acute lumbosacral PR in HIV-infected patients must be considered a syndrome with different causes. CMV and M. tuberculosis infections were the most frequent causative agents in our series (41% and 18% of the cases, respectively). Early empirical therapy is often necessary as definite diagnosis may be delayed or never achieved. Our experience suggests that, at least in our milieu, anti-tuberculous drugs should be considered in some cases together with ganciclovir or foscarnet in the empirical therapy for PR in HIV-infected patients.

[Community-acquired pneumonia: prospective study of 101 adult, immunocompetent patients for 1 year]

BACKGROUND

A one year prospective study was carried out to assess the etiology of community-acquired pneumonia (CAP), and also to know the incidence, characteristics and evolution of infection by Chlamydia pneumoniae; and the effectiveness of DNA probes in CAP due to Mycoplasma pneumoniae and Legionella.

METHODS

One hundred and ten patients with a diagnosis of CAP in the emergency department were studied. Serologic studies were performed, and also tests commonly used for the diagnosis of respiratory tract pathogens in respiratory samples, including serology and culture of Chlamydia pneumoniae and DNA probes for Mycoplasma pneumoniae and Legionella.

RESULTS

In 72 cases (71.3%) some pathogen was found and in 5 cases more than one microorganism was involved. The etiology was bacterial in 31% of the cases, with S. pneumoniae being the most frequent (19 cases). Forty percent of the cases were "atypical" pneumonias with 33 cases of M. pneumoniae and 5 by Chlamydia pneumoniae. Diagnostic data of viral pneumonia were found in 2 cases. DNA probes were not useful for the diagnosis of pneumonia by Legionella pneumophila and had low effectiveness (31.8%) in Mycoplasma pneumoniae CAP.

CONCLUSIONS

a) M. pneumoniae was the most frequent pathogen (33%). b) DNA probes for M. pneumoniae had low sensitivity in sputum (31.8%) and none in pharyngeal exudate. c) Acute infection by C. pneumoniae was diagnosed in 5 cases. Previous data of infection were recorded in 60.4% of the patients. d) Bacterial pneumonia (31%) was underestimated due to a low rate of bacteremic cases (7.9%) and the low number of positive cultures with definitive diagnostic value. e) The evolution was good except in two cases (death due to staphylococcal pneumonia with alcohol withdrawal syndrome and multiorganic failure by disseminated chicken-pox).

Pneumocystis carinii pneumonia in HIV-infected patients: diagnostic yield of induced sputum and immunofluorescent stain with monoclonal antibodies

The purpose of this study was to evaluate the diagnostic yield of induced sputum (IS), assessing the reliability of indirect immunofluorescent stain with monoclonal antibodies (IFMoAb) and methenamine silver (Met-Ag) and analysing factors likely to influence the sensitivity of these techniques. An analysis was prospectively carried out on IS specimens collected from 61 human immunodeficiency virus (HIV)-infected patients during 69 episodes of suspected Pneumocystis carinii pneumonia. Ultrasonic nebulizers with hypertonic 2% saline were used. IFMoAb to P. carinii and Met-Ag were performed after cytocentrifugation of the specimen. Results were compared with those of bronchoalveolar lavage (BAL) with/without transbronchial biopsy (TBB), performed not more than seven days after induction of sputum. P. carinii pneumonia was confirmed in 32 episodes, of which IS was diagnostic in 23. The sensitivity of the staining procedures was 69% for IFMoAb, and 28% for Met-Ag. The three episodes of P. carinii pneumonia in patients on oral chemoprophylaxis yielded negative IS results; in contrast, IS was negative in only 6 of the 29 cases not receiving chemoprophylaxis. IS is a non-aggressive procedure that diagnosed P. carinii pneumonia in 72% of our cases. The yield increased significantly when IFMoAb was used in patients not receiving oral chemoprophylaxis.

[Pneumococcal endocarditis: apropos of 3 cases]

Streptococcus pneumoniae is a common pathogen in man. The incidence of endocarditis caused by this organism has diminished in the last decades, and now it is an exceptional condition. Three patients with pneumococcal endocarditis are reported. Their clinical and epidemiological features are reviewed and compared with previous literature. One of the instances was due to a strain with high resistance to penicillin; vancomycin was a valid alternative therapy.