Incidence Of Pneumocystis Carinii Pneumonia Research Articles

Incidence of Pneumocystis carinii pneumonia in treatment-naïve HIV-infected adult patients receiving primary Pneumocystis carinii pneumonia chemoprophylaxis with 48-week antiretroviral initiation

Incidence of Pneumocystis carinii pneumonia in treatment-naïve HIV-infected adult patients receiving primary Pneumocystis carinii pneumonia chemoprophylaxis with 48-week antiretroviral initiation

Clinical diagnosis and treatment of pneumocystis carinii pneumonia in children with blood disease

Pneumocystis carinii pneumonia (PCP) is a life-threatening opportunistic infection in immunocompromised patients.In recent years, the incidence of PCP has increased significantly since the application of combined chemotherapy, immunosuppressant and hematopoietic stem cell transplantation, with improved long-term efficacy in pe-diatric blood disease.This review summarizes the clinical manifestations, laboratory and radiological examination, diagnosis, treatment and prophylaxis.Compound Sulfamethoxazole is the most commonly used agent for prophylaxis.It is recommended for the high-risk patients to take PCP prophylaxis, which could decrease the incidence of PCP. Key words: Pneumocystis; Pneumonia; Prophylaxis; Child

Pneumocystis Pneumonia in Solid Organ Transplant Recipients

Pneumocystis spp. are thought to be ubiquitous in nature with serologic studies suggesting that exposure occurs commonly in childhood (2). Symptomatic Pneumocystis pneumonia (PCP) is generally limited to individuals with immune deficits. Animal models demonstrate both de novo infection via airborne transmission and by reactivation of previously established latent infections (3). Clusters of infection have been described in medical facilities among renal transplant recipients (4–6). Based on studies before routine implementation of prophylaxis, the overall incidence of infection among solid organ transplant recipients varied mostly in the range of 5–15% depending on organ type, transplant center and immunosuppressive regimens (7). In older reports before implementation of extensive prophylaxis, the attack rate appeared highest in lung and combined heart–lung transplant recipients with an overall incidence ranging from less than 10% to just over 40%. Among liver and kidney transplant recipients, the incidence appeared to be less at anywhere from close to 2% to 10–15% (1). The incidence overall may be decreasing with reduction in the routine use of corticosteroids in organ transplantation and the adoption of effective prophylactic measures. Table 1 outlines some risk factors for PCP among non-HIV-infected individuals.

Lymphocyte Depletion in Breast Cancer Patients Receiving Adjuvant Dose-Dense Chemotherapy.

Background: Dose-intense chemotherapy regimens are reportedly known to cause lymphocyte depletion. Increased incidence of pneumocystis carinii pneumonia (PCP) was reported in patients with diffuse large B-cell lymphoma treated with dose-dense CHOP plus rituximab. No studies have been conducted to date to assess the lymphocyte depletion in breast cancer patients receiving dose-dense chemotherapy. We retrospectively compared the lymphocyte recovery in breast cancer patients treated with dose-dense (q 2 weekly) doxorubicin/cyclophosphamide (AC) to those treated with AC in a non-dose-dense (q 3 weekly) method.Methods: Newly diagnosed breast cancer patients treated at our institution between 2002 and 2006, with at least 4 cycles of AC in an adjuvant/neoadjuvant setting were analyzed. Complete blood count (CBC) with differential including white blood cells (WBC), hemoglobin (Hg), platelets, neutrophils, lymphocytes, monocytes, eosinophils and basophils were collected at baseline before the start of chemotherapy, before each ensuing cycle of chemotherapy and after the 4th cycle of chemotherapy.Results: A total of 103 patients were analyzed of which 50 patients were treated with dose-dense AC and 53 patients with non-dose-dense AC. After the 4th cycle of chemotherapy, mean absolute lymphocyte count (ALC) was 780 ± 440/mm3 in the dose-dense group versus 950 ± 600/mm3 in the non-dose-dense group (p = .4617), mean Hg was 11.4 ± 1.2g/dL in the dose-dense group versus 12.2 ± 0.9g/dL in the non-dose-dense group, mean platelet count was 241 ± 85 x109/L in the dose-dense group versus 391 ± 128 x109/L in the non-dose-dense group. No opportunistic infections were noted in the dose-dense group.Conclusions: These results demonstrate that dose-dense AC results in slightly increased lymphocyte depletion compared to non-dose-dense AC, not statistically significant. This didn't result in any increased infectious complications. In both groups, baseline ALC was significantly positively related to ALC after the 4th cycle of chemotherapy (p < 0.0001). Dose-dense AC did result in statistically significant platelet depletion.

Clinical features of HIV/AIDS patients presenting to an inner city clinic in Ho Chi Minh City, Vietnam

An outpatient HIV clinic was opened in March 2005 in Binh Thanh District, a poor section of Ho Chi Minh City, Vietnam. Over 1500 patients were seen in the first year. The average age of patients was 27 years. Men represented 77% of the clinic population, women, 23% and children under the age of 16 years of age, 5% of the population. The most common risk factor among men was being an injecting drug user (IDU), 76%, and among women, being married to an IDU HIV-positive man, 35%. Physical signs of disease were uncommon: lymphadenopathy in 24% and hepatomegaly and splenomegaly in 4% and 3%, respectively. Men and women were anaemic at presentation, with a mean haemoglobin of 11.9 g/dL and 11.1 g/dL, respectively. An overwhelming majority of patients had profound immunodeficiency. The mean CD4+ cell count was 164 cells/mL and the median was 69 cells/mL. No correlation was found between the World Health Organization's stage of disease and the CD4+ cell count. Thus, the former is a poor predictor of immunity in this population. Data regarding opportunistic infections diagnosed at the first visit were studied. Candidiasis of the oral pharynx, oesophagus or vagina was found in 34.5% of the patients, and pulmonary and extrapulmonary tuberculosis was found in 32% of the patients. Pneumocystis carinii pneumonia (PCP) was diagnosed in only 3% of the patients. Cotrimoxazole prophylaxis is advocated for HIV-infected Vietnamese, but the incidence of PCP is negligible and resources could be spent elsewhere. The various opportunistic infections seen in this resource-poor clinic setting is likely to be a pattern of presentation of HIV-infected Vietnamese for some time to come.

Audit of pneumocystis pneumonia in patients seen by the Christchurch Hospital rheumatology service over a 5‐year period

The aim of the study was to review all cases of Pneumocystis carinii pneumonia (PCP) in patients seen by the Christchurch Hospital Rheumatology service over a 5-year period and to determine the annual incidence of PCP. The Canterbury Health Laboratory database was searched for rheumatology patients testing positive for PCP from 31 December 2000 to 31 December 2005. The rheumatology database was then searched to identify patients receiving the same immunosuppressant medication as those who developed PCP to determine the annual incidence of PCP in this group. Four rheumatology patients were diagnosed with PCP during the 5-year period. Two were receiving oral methotrexate (MTX) for rheumatoid arthritis and two were receiving cyclophosphamide (CYC), one each for Wegener's granulomatosis and dermatomyositis. None of the four cases was receiving PCP chemoprophylaxis. Five hundred and forty-seven patients commenced MTX over the same 5-year period and 47 commenced CYC. Only 14 of 47 (29.7%) CYC-treated patients received PCP prophylaxis. The annual incidence of PCP was 0.17% (95% confidence interval (CI) 0.02-0.63) and 5.33% (95%CI 0.65-19.24) in patients prescribed MTX and CYC, respectively. For the 33 patients receiving CYC without concomitant PCP prophylaxis the annual incidence was 9.50% (95%CI 1.15-34.33). In our study the annual incidence of PCP in patients taking MTX was low and would not support the use of routine PCP chemoprophylaxis. In patients receiving CYC without concomitant PCP chemoprophylaxis the annual incidence of PCP was higher although the number of cases was small. Given the high morbidity and mortality in this group, PCP chemoprophylaxis should be considered.

Prophylaxis of Pneumocystis carinii pneumonia with atovaquone in children with leukemia

Despite extensive studies of atovaquone in human immunodeficiency virus (HIV)-infected patients, there is little information about its efficacy as a prophylactic agent for Pneumocystis carinii pneumonia (PCP) in pediatric patients with cancer. Therefore, a retrospective analysis was conducted to determine the incidence of PCP in pediatric patients who received prophylactic atovaquone during treatment for acute leukemia. We reviewed the medical records of all patients treated at our institution for acute lymphoblastic leukemia or acute myeloid leukemia between 1994 and 2004. Only patients who were intolerant of trimethoprim-sulfamethoxazole (TMP-SMZ) and received atovaquone prophylaxis were included in the analysis. Eighty-six patients were unable to tolerate TMP-SMZ and received daily atovaquone for PCP prophylaxis. PCP was not diagnosed in any patient who received atovaquone prophylaxis: the upper limit of the 95% confidence interval (CI) was 1.74 per 100 person-years. Atovaquone is an efficacious alternative for PCP prophylaxis in pediatric patients who have leukemia and are intolerant of TMP-SMZ.

Mycophenolate mofetil may protect against Pneumocystis carinii pneumonia in renal transplanted patients

Pneumocystis carinii pneumonia (PCP) is usually prevented in transplanted patients by prophylactic trimethoprim-sulfamethoxazol (TMS). Mycophenolate mofetil (MMF) has been shown to have a strong protective effect against PCP in rats. This effect is also suggested in humans by the absence of PCP in patients receiving MMF. After January 1998 MMF has been used with no TMS prophylaxis in renal transplanted patients. In azathioprine (AZA) treated patients TMS prophylaxis was maintained. The incidence of PCP was analyzed in both groups. Data were collected in order to have a minimum 6-month follow-up. Two hundred and seventy-two patients were eligible for analysis. No PCP occurred either in patients under MMF without TMS prophylaxis nor in patients under AZA. MMF may have an effective protective role against PCP as no patient under MMF, despite not receiving TMS coverage, developed PCP. A larger, controlled, trial is warranted to consolidate this information.

Chronic Steroid and Immunosuppressant Therapy in Children

1. Abraham Gedalia, MD* 2. Avinash K. Shetty, MD† 1. *Professor of Pediatrics and Head, Division of Pediatric Rheumatology, Louisiana State University Health Sciences Center and Children’s Hospital, New Orleans, La 2. †Assistant Professor of Pediatrics, Division of Infectious Diseases, Wake Forest University School of Medicine, Winston-Salem, NC After completing this article, readers should be able to: 1. Describe the mechanism of action, list indications for therapy, and discuss the adverse effects of corticosteroid therapy in children. 2. Delineate the mechanism of action and adverse effects of other immunosuppressive agents frequently used in children. 3. Discuss the immunization recommendations for children receiving immunosuppressive therapy. Corticosteroids are the most potent, naturally occurring anti-inflammatory agents known. Despite more than 50 years of controversy regarding their clinical use, risks, and benefits, corticosteroids still play a significant role in the management of a wide variety of diseases. They have reduced the morbidity and mortality significantly in children afflicted with leukemia, asthma, rheumatic diseases, and other inflammatory disorders. In this article, we review the mechanism of action, clinical indications, and adverse effects of corticosteroids and other commonly used immunosuppressive agents in children. We also review immunization recommendations for children receiving immunosuppressive therapy. Corticosteroids have a diverse role in the human body at the physiologic level. The actions of glucocorticoids include: 1) negative feedback modulation of corticotropin-releasing factor and adrenocorticotropic hormone; 2) regulation of blood glucose and liver glycogen levels; 3) maintenance of water-electrolyte homeostasis; 4) influence on the metabolism of protein, fat, and purine; 5) maintenance of normal cardiovascular, central nervous system (CNS), and renal functions; 6) influence on circadian rhythmicity; 7) maintenance of bone and muscle work capacity; and 8) protection of the body against moderate stress. Glucocorticoids also have pharmacologic properties that are used in the treatment of rheumatic and other inflammatory diseases. Cellular receptors for cortisol are ubiquitous in cell cytoplasm, reflecting the critical role of the hormone in maintaining cell homeostasis. The type 1 glucocorticoid receptor, known as the mineralocorticoid receptor, has high affinity for aldosterone. The type 2 glucocorticoid receptor, known as the glucocorticoid receptor, exhibits strong affinity for dexamethasone. Although …

Prospective cohort study showing changes in the monthly incidence of Pneumocystis carinii pneumonia

Pneumocystis carinii pneumonia (PCP) remains a serious opportunistic infection in HIV infected individuals. Seasonal changes in climate are associated with changes within individual susceptibility to infection. The possibility of monthly...

Comparison of Clinical Characteristics of Pneumocystis Carinii Pneumonia Between HIV Infected and Non-infected Persons

Background : Pneumocystis carinii pneumonia (PCP) is one of the most common cause of infection in patients with HIV infection. Recently, the incidence of PCP have been increasing in immunocompromised hosts without HIV infection. We compared the clinical characteristics of PCP between HIV infected and non-infected persons. Patients and Methods : We retrospectively reviewed the charts of 25 patients diagnosed as PCP from 1996 to 2002. Age, sex, underlying conditions, use of immunosuppressants, clinical courses, laboratory findings, treatment and prognosis were compared between HIV infected and non-infected persons. Results : Twenty-five patients with PCP were identified. 16 were HIV infected, and 9 were HIV non-infected. The mean age of overall patients was 43.4±13.2 years. Underlying conditions in HIV non-infected persons were hematologic malignancy (7 cases), solid organ transplant (1 case), and autoimmune disease (1 case). Seven cases (77.8%) of HIV non-infected persons had a history of steroid use. Mean duration of symptoms was longer in HIV infected persons than in HIV non-infected persons, but it was not statistically significant. PaO2 was lower in HIV infected persons (61.2±16.9 mmHg vs.

Granulomatous Pneumocystis carinii pneumonia in patients with malignancy

Background: A review was undertaken of the clinical features and results of diagnostic tests in non-HIV infected patients who developed granulomatous Pneumocystis carinii pneumonia (PCP). Methods: A retrospective review was...

Incidence of primary opportunistic infections in two human immunodeficiency virus-infected French clinical cohorts.

Clinical guidelines for the prevention of opportunistic infections in human immunodeficiency virus (HIV)-infected individuals have been developed on the basis of natural history data collected in the USA. The objective of this study was to estimate the incidence of primary opportunistic infections in HIV-infected individuals in geographically distinct cohorts in France. We conducted our study on 2664 HIV-infected patients from the Tourcoing AIDS Reference Centre and the hospital-based information system of the Groupe d'Epidémiologie Clinique du SIDA en Aquitaine enrolled from January 1987 to September 1995 and followed through December 1995. We estimated: (1) CD4-adjusted incidence rates of seven primary opportunistic infections in the absence of prophylaxis for that specific infection or any antiretroviral drugs other than zidovudine; and (2) CD4 lymphocyte count decline. The highest incidence rates for all opportunistic infections studied occurred in patients with CD4 counts < 200/microl. With CD4 counts < 50/microl, the most common opportunistic infections were toxoplasmic encephalitis (12.6 per 100 person-years) and Pneumocystis carinii pneumonia (11.4 per 100 person-years). Mycobacterium tuberculosis was the least common opportunistic infection (< 5.0/100 person-years). Even with CD4 counts > 300/microl, cases of Pneumocystis carinii pneumonia and toxoplasmic encephalitis were reported. The mean CD4 lymphocyte decline per month was 4.6 cells/microl. There was a significant association between HIV risk behaviour and the incidence of cytomegalovirus infection, between calendar year and the incidence of Pneumocystis carinii pneumonia, toxoplasmic encephalitis and Candida esophagitis, and between geographical area and the incidence of Pneumocystis carinii pneumonia and cytomegalovirus infection. Geographical differences exist in the incidence of HIV-related opportunistic infections. These results can be used to define local priorities for prophylaxis of opportunistic infections.

Pneumocystis carinii pneumonia during maintenance treatment of childhood acute lymphoblastic leukemia.

Pneumocystis carinii pneumonia (PCP) is a wellknown risk among patients with deficient T-cell function such as children treated for acute lymphoblastic leukemia (ALL). The purpose of this study was to estimate the risk for PCP during maintenance treatment (MT) to identify patients at risk who could benefit from prophylaxis. We registered all episodes of PCP during MT in 71 children diagnosed between January 1992 and June 1997 with non-B-cell ALL at The Copenhagen University Hospital. Sulphametoxazole and trimetroprim (SMX/TMP) prophylaxis against PCP was given during induction and consolidation therapy but stopped prior to MT with oral methotrexate/6-mercaptopurine. Patients with standard (SR), intermediate (IR), and high risk (HR) ALL started MT at 3, 8, and 15 months from diagnosis, respectively. The HR group had a cumulated risk of 70% for developing PCP, whereas the risk for PCP in children with IR and the SR was 11 and 8%, respectively (P < 0.0001). All but one of these 13 cases of PCP occurred within 8 months after cessation of SMX-TMP prophylaxis. The higher incidence of PCP among HR compared to non-HR patients following cessation of SMX/TMP prophylaxis probably reflects the significantly longer T-cell suppressive consolidation therapy in this group. The very low incidence of PCP during the later part of MT emphasizes that methotrexate/6-mercaptopurine MT have more impact on B-cell than on T-cell function. TMP/SMX prophylaxis should be recommended for all children treated for ALL.

Pulmonary infiltrates in HIV-infected patients in the highly active antiretroviral therapy era in Spain.

To study the incidence, etiology, and outcome of pulmonary infiltrates (PIs) in HIV-infected patients and to evaluate the yield of diagnostic procedures. Prospective observational study of consecutive hospital admissions. Tertiary hospital. HIV-infected patients with new-onset radiologic PIs from April 1998 to March 1999. The study protocol included chest radiography, blood and sputum cultures, serologic testing for "atypical" causes of pneumonia, testing for Legionella urinary antigen, testing for cytomegalovirus antigenemia, and bronchoscopy in case of diffuse or progressive PIs. One hundred two episodes in 92 patients were recorded. The incidence of PIs was 18 episodes per 100 hospital admission-years (95% confidence interval [CI]: 15-21). An etiologic diagnosis was achieved in 62 cases (61%). Bacterial pneumonia (BP), Pneumocystis carinii pneumonia (PCP), and mycobacteriosis were the main diagnoses. The incidences of BP and mycobacteriosis were not statistically different in highly active antiretroviral therapy (HAART) versus non-HAART patients. The incidence of PCP was lower in those receiving HAART (p =.011), however. Nine patients died (10%). Independent factors associated with higher mortality were mechanical ventilation (odds ratio [OR] = 83; CI: 4.2-1,682), age >50 years (OR = 23; CI: 2-283), and not having an etiologic diagnosis (OR = 22; CI: 1.6-293). Pulmonary infiltrates are still a frequent cause of hospital admission in the HAART era, and BP is the main etiology. There was no difference in the rate of BP and mycobacteriosis in HAART and non-HAART patients. Not having an etiologic diagnosis is an independent factor associated with mortality.

Risk Factors for Prophylaxis Failure in Patients Receiving Aerosol Pentamidine for Pneumocystis carinii Pneumonia Prophylaxis

The purposes of this study were (1) to determine the incidence of prophylaxis failure in HIV-infected patients receiving aerosol pentamidine (AP) for Pneumocystis carinii pneumonia (PCP) prophylaxis, and (2) to identify risk factors for PCP prophylaxis failure. In Ontario, Canada, AP has been made available for outpatient PCP prophylaxis through a centralized government program, the Ontario Drug Distribution and Monitoring Program. Data from this administrative observational database were extracted for 2,227 patients who received AP between May 1989 and December 1998. The incidence of breakthrough PCP (BPCP) was calculated from the database. A Cox regression model with time-varying covariates was created to examine factors associated with BPCP. The follow-up time was divided into three eras: 1989 to 1991, 1992 to 1994, and 1995 to 1998. These eras were meant to reflect major changes in antiretroviral medication regimens. The overall risk of BPCP was 16.2% over a mean follow-up of 1.67 years. The overall BPCP rate was 9.7/100 patient-years, with rates of 8.8/100, 13.1/100, and 6.3/100 patient-years in each of the three treatment eras. In the multivariate analysis, significant risk factors for prophylaxis failure were low CD4 count, previous diagnosis of PCP, history of AIDS-defining conditions other than PCP, and antiretroviral treatment era defined above. The overall rate of PCP prophylaxis failure has decreased significantly after 1995, coincident with the era of highly active antiretroviral therapies. Initiation of PCP prophylaxis remains necessary in patients with risk factors.

Pneumocystis carinii pneumonia: a review of current issues in diagnosis and management.

Pneumocystis carinii pneumonia: a review of current issues in diagnosis and management.

Respiratory diseases in the first year of life in children born to HIV-1-infected women.

Our objective was to describe the respiratory complications, clinical findings, and chest radiographic changes in the first year of life in infected and uninfected children born to HIV-1-infected women. We prospectively followed a cohort of 600 infants born to HIV-1-infected women from birth to 12 months in a multicenter study. Of these, 93 infants (15.5%) were HIV-1-infected, 463 were uninfected, and 44 were of unknown status prior to death or loss to follow-up. The cumulative incidence ( +/- SE) of an initial pneumonia episode at 12 months was 24.1 +/- 4.7% in HIV-1-infected children compared to 1.4 +/- 0.6% in HIV-1-uninfected children (P < 0.001). The rate of Pneumocystis carinii pneumonia (PCP) was 9.5 per 100 child-years. The HIV-1 RNA load was not higher in the group that developed pneumonia in the first year vs. those who did not. Children who developed lower respiratory tract infections or PCP had increased rates of decline of CD4 cell counts during the first 6 months of life. Lower maternal CD4 cell counts were associated with higher rates of pneumonia, and upper and lower respiratory tract infections. The rates of upper respiratory tract infection and bronchiolitis/reactive airway disease in infected children were not significantly different than in uninfected children. At 12 months, significantly more HIV-1-infected than uninfected children had tachypnea and chest radiographs with nodular and reticular densities. There was no relationship between cytomegalovirus infection in the first year of life and radiographic changes or occurrences of pneumonia. In conclusion, despite a low incidence of PCP, rates of pneumonia remain high in HIV-infected children in the first year of life. The incidence of pneumonia in uninfected infants born to HIV-1-infected mothers is low. Chest X-ray abnormalities and tachypnea suggest that subacute disease is present in infected infants. Further follow-up is warranted to determine its nature.

The Changing Pattern of AIDS-defining Illnesses with the Introduction of Highly Active Antiretroviral Therapy (HAART)in a London Clinic

Objectives:To quantify the progressive impact of combination antiretroviral therapy (ART) on the incidence of AIDS-defining illnesses (ADIs) over a 9-year period.Methods: Retrospective cohort study. Eligible patients were 1538 AIDS-free, HIV-1-positive patients attending a large HIV clinic in west London who were at risk of developing AIDS because their CD4 count had declined to O350e106/1 cells during the period 1 January 1990 and 31 December 1998. Incidence rates for the 12 most frequent ADIs were compared for two time periods, 1990–1995 (pre-HAART) and 1996–1998 (post-HAART), using Poisson regression methods. Multivariate Poisson regression models were used to examine the contribution of ART and HAART to any observed temporal trends in incidence rates.Results: After a median follow-up of 35 months, 450 (29%) patients had developed AIDS. Between the two time periods there was a significant decrease in the incidence ofPneumocystis carinii pneumonia (PCP) by 35% (4.11 per 100 person-years in 1990–1995 vs. 2.67 in 1996–1998;P= 0.007), Kaposi's sarcoma by 34% (3.27 vs. 2.17;P= 0.022) and cryptosporidiosis by 60% (0.76 vs. 0.31;P= 0.029). A non-significant reduction in incidence was observed for cryptococcosis by 45% (0.81 vs. 0.45;P= 0.11), oesophageal candidiasis by 29% (3.34 vs. 2.39;P= 0.053) and mycobacterium avium complex by 18% (1.58 vs. 1.29;P= 0.4), and a non-significant increase was observed for tuberculosis by 17% (0.62 vs. 0.73;P= 0.66) and non-Hodgkins lymphoma (NHL) by 51% (0.43 vs. 0.65;P= 0.31). The incidence of cerebral toxoplasmosis, cytomegalovirus, recurrent bacterial chest infections and dementia remained stable. There was a clear stepwise reduction in the incidence of PCP, Kaposi's sarcoma and cryptosporidiosis with the use of non-H AART and HAART regimens relative to no ART. In a multivariate analysis, the use of ART and HAART explained the progressive decrease in incidence of PCP and Kaposi's sarcoma.Conclusions:The incidence of most ADIs has decreased over the last 9 years. The striking reduction in the inci-dence of PCP and Kaposi's sarcoma since 1996 can be attributed to the use of combination ART and particularly HAART. The non-significant increase in the incidence of NHL and tuberculosis needs confirmation in other patient cohorts.

Epidemiology of Pneumocystis carinii pneumonia in an era of effective prophylaxis: the relative contribution of non-adherence and drug failure

To determine the relative contribution of patient non-adherence, provider failure to prescribe prophylaxis, and drug failure to the continued occurrence of Pneumocystis carinii pneumonia (PCP), and to determine correlates of non-adherence. Retrospective case-control study. Patients with confirmed or presumptive PCP from May 1995 to September 1997 who had at least 6 months of prior HIV care (cases) were compared to controls matched for initial CD4 cell count and date of initial HIV care. The incidence of PCP declined by 85% in the 28 months of the study. Of the 118 cases of PCP identified, 59 (50%) were in HIV care for > 6 months prior to PCP diagnosis. In a multivariate logistic regression model, risk factors for PCP among patients in HIV care were patient non-adherence [odds ratio (OR), 12.4; 95% confidence interval (CI), 6.4-23.5], use of prophylaxis other than trimethoprim-sulfamethoxazole (OR, 27.0; 95% CI, 13.8-52.9), and absence of antiretroviral use (OR, 7.5; 95% CI, 4.5-12.5). Provider non-adherence occurred in one out of 59 cases (2%), and five out of 106 controls (5%). Of the patients who developed PCP on prophylaxis, 18 cases (30%) appeared due to drug failure; there were no cases of apparent drug failure among patients on trimethoprim-sulfamethoxazole. In multivariate analysis, non-adherence was more common among patients of non-white race, those with a history of injecting drug use, and those with active substance abuse or psychiatric illness. Patient non-adherence was the most common reason for the occurrence of PCP among patients in HIV care; provider non-adherence was uncommon. Drug failure occurred only among patients on prophylaxis other than trimethoprim-sulfamethoxazole.