Multisystem organ failure predicts mortality of ICU patients with acute respiratory failure secondary to AIDS-related PCP

Evolutionary warning system for COVID-19 severity: Colony predation algorithm enhanced extreme learning machine

Coronavirus Disease 2019 (COVID-19) was distributed globally at the end of December 2019 due to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Early diagnosis and successful COVID-19 assessment are missing, clinical care is ineffective, and deaths are high. In this study, we investigate whether the level of biochemical indicators helps to discriminate and classify the severity of the COVID-19 using the machine learning method. This research creates an efficient intelligence method for the diagnosis of COVID-19 from the perspective of biochemical indexes. The framework is proposed by integrating an enhanced new stochastic called the colony predation algorithm (CPA) with a kernel extreme learning machine (KELM), abbreviated as ECPA-KELM. The core feature of the approach is the ECPA algorithm which incorporates the two main operators that have been abstained from the grey wolf optimizer and moth-flame optimizer to improve and restore the CPA research functions and are simultaneously used to optimize the parameters and to select features for KELM. The ECPA output is checked thoroughly using IEEE CEC2017 benchmark to verify the capacity of the proposed methodology. Finally, in the diagnosis of COVID-19 using biochemical indexes, the designed ECPA-KELM model and other competing KELM models based on other optimization are used. Checking statistical results will display improved predictive properties for all metrics and higher stability. ECPA-KELM can also be used to discriminate and classify the severity of the COVID-19 as a possible computer-aided method and provide effective early warning for the treatment and diagnosis of COVID-19.

Developing and validating COVID-19 adverse outcome risk prediction models from a bi-national European cohort of 5594 patients

Patients with severe COVID-19 have overwhelmed healthcare systems worldwide. We hypothesized that machine learning (ML) models could be used to predict risks at different stages of management and thereby provide insights into drivers and prognostic markers of disease progression and death. From a cohort of approx. 2.6 million citizens in Denmark, SARS-CoV-2 PCR tests were performed on subjects suspected for COVID-19 disease; 3944 cases had at least one positive test and were subjected to further analysis. SARS-CoV-2 positive cases from the United Kingdom Biobank was used for external validation. The ML models predicted the risk of death (Receiver Operation Characteristics—Area Under the Curve, ROC-AUC) of 0.906 at diagnosis, 0.818, at hospital admission and 0.721 at Intensive Care Unit (ICU) admission. Similar metrics were achieved for predicted risks of hospital and ICU admission and use of mechanical ventilation. Common risk factors, included age, body mass index and hypertension, although the top risk features shifted towards markers of shock and organ dysfunction in ICU patients. The external validation indicated fair predictive performance for mortality prediction, but suboptimal performance for predicting ICU admission. ML may be used to identify drivers of progression to more severe disease and for prognostication patients in patients with COVID-19. We provide access to an online risk calculator based on these findings.

Micoses sistêmicas: fatores associados ao óbito em pacientes com infecção pelo vírus da imunodeficiência humana, Cuiabá, Estado de Mato Grosso, 2005-2008

A prevalência de micose sistêmica entre 1.300 pacientes portadores de HIV/Aids de Cuiabá, Mato Grosso foi de 4,6%, no período de 2005-2008. As espécies de fungos isoladas foram o Cryptococcus neoformans (50%), Cryptococcus gattii (1,6%), Cryptococcus spp (6,6%), Histoplasma capsulatum (38,3%) e Paracoccidioides brasiliensis (3,3%). Óbito foi registrado em 32 (53,3%) pacientes, sendo a criptococose a principal causa. A contagem de linfócitos T CD4+ foi baixa e semelhante entre os pacientes que sobreviveram ou faleceram por micose sistêmica. O etilismo (OR:8,2; IC95%: 1,4-62,1; p=0,005) e o nível médio de desidrogenase lática [758 (182) U/L vs 416 (268) U/L; p<0,001] foram as características independentemente associadas ao óbito dos pacientes do estudo. Os resultados mostram alta letalidade por micoses sistêmicas em pacientes portadores de HIV/Aids de Cuiabá e sugerem que características clínico-laboratoriais tais como o etilismo e a elevação precoce da desidrogenase lática podem ser fatores relacionados ao pior prognóstico nessas condições.

Hematological and biochemical factors predicting SARS fatality in Taiwan

BACKGROUND/PURPOSE

Severe acute respiratory syndrome (SARS) has a high fatality rate worldwide. We examined the epidemiologic and clinical factors associated with death for all laboratory-confirmed SARS patients in Taiwan.

METHODS

Using initial data in medical records reported by hospitals to the Center for Disease Control in Taiwan, we analyzed whether hematological, biochemical and arterial blood gas measures could predict fatality in 346 SARS patients.

RESULTS

Both fatalities (n = 73; 21.1%) and survivors had elevated plasma concentration of initial C-reactive protein (CRP), but the mean CRP concentration was higher in fatalities (47.7 +/- 43.3 mg/L) than in survivors (24.6 +/- 28.2 mg/L). Initial lymphocyte counts were low in both fatalities (814 +/- 378/microL) and survivors (1019 +/- 480/microL). After controlling for age and sex, multiple logistic regression analysis showed that hematological factors significantly associated with fatality included initial neutrophil count > 7000/microL (odds ratio [OR] = 6.4), initial CRP concentration > 47.5 mg/L (OR = 5.8) and lactic acid dehydrogenase (LDH) > 593.5 IU/L (OR = 4.2). Factors significantly associated with initial CRP concentration > 47.5 mg/L included dyspnea (OR = 4.3), red blood cell count < 4.1 x 106/microL (OR = 4.3) and serum aspartate aminotransferase > 57 IU/L (OR = 3.1).

CONCLUSION

Initial neutrophil count, CRP and LDH levels are important predictors of mortality from SARS.

Improved survival for HIV infected patients with severe Pneumocystis jirovecii pneumonia is independent of highly active antiretroviral therapy

BACKGROUND

Despite a decline in incidence of Pneumocystis jirovecii pneumonia (PCP), severe PCP continues to be a common cause of admission to the intensive care unit (ICU) where mortality remains high. A study was undertaken to examine the outcome from intensive care for patients with PCP and to identify prognostic factors.

METHODS

A retrospective cohort study was conducted of HIV infected adults admitted to a university affiliated hospital ICU between November 1990 and October 2005. Case note review collected information on demographic variables, use of prophylaxis and highly active antiretroviral therapy (HAART), and hospital course. The main outcome was 1 month mortality, either on the ICU or in hospital.

RESULTS

Fifty nine patients were admitted to the ICU on 60 occasions. Thirty four patients (57%) required mechanical ventilation. Overall mortality was 53%. No patient received HAART before or during ICU admission. Multivariate analysis showed that the factors associated with mortality were the year of diagnosis (before mid 1996 (mortality 71%) compared with later (mortality 34%; p = 0.008)), age (p = 0.016), and the need for mechanical ventilation and/or development of pneumothorax (p = 0.031). Mortality was not associated with sex, ethnicity, prior receipt of sulpha prophylaxis, haemoglobin, serum albumin, CD4 count, PaO2, A-aO2 gradient, co-pathology in bronchoscopic lavage fluid, medical co-morbidity, APACHE II score, or duration of mechanical ventilation.

CONCLUSIONS

Observed improved outcomes from severe PCP for patients admitted to the ICU occurred in the absence of intervention with HAART and probably reflect general improvements in ICU management of respiratory failure and ARDS rather than improvements in the management of PCP.

Electrocardiogram in Pneumocystis carinii pneumonia: can it be used as a prognostic variable?

OBJECTIVE

Many prognostic variables have been studied in patients with Pneumocystis carinii pneumonia and acquired immunodeficiency syndrome (AIDS). The role of the electrocardiogram in this setting has not been previously evaluated. We analyzed the admission electrocardiogram in patients with Pneumocystis carinii pneumonia and AIDS in an attempt to identify electrocardiogram findings that could be associated with adverse clinical outcomes and worse prognostic variables.

DESIGN

A retrospective medical chart review.

SETTING

All confirmed cases of Pneumocystis carinii pneumonia in patients positive for human immunodeficiency virus admitted to Albert Einstein Medical Center from 1994 to 2000.

METHODS

Patients were assigned increasing severity ranks based on the findings on the admission electrocardiogram (normal sinus rhythm, sinus tachycardia, and right ventricular strain pattern). Data were extracted regarding study outcomes (admission to intensive care unit, mechanical ventilation, and hospital mortality) and prognostic variables.

MAIN RESULTS

Of the 40 study patients, 14 (35%) had normal sinus rhythm, 15 (37.5%) had sinus tachycardia, and 11 (27.5%) presented with signs of right ventricular strain. The number of admissions to the intensive care unit, use of mechanical ventilation, and hospital mortality rate all increased with the severity of the electrocardiogram findings (p < or =.03). The serum lactate dehydrogenase concentrations and the alveolar-arterial oxygen gradient both increased with the severity of the electrocardiogram findings (p < or =.02).

CONCLUSION

Electrocardiogram findings of sinus tachycardia and right heart strain are common in Pneumocystis carinii pneumonia. These findings are associated with adverse clinical outcomes as well as worsening of prognostic variables. The electrocardiogram may be useful in predicting outcome in patients with Pneumocystis carinii pneumonia.

Acute respiratory failure in the United States: incidence and 31-day survival

STUDY OBJECTIVES

To estimate the incidence of acute respiratory failure (ARF) in the United States and to analyze 31-day hospital mortality among a cohort of patients with ARF.

DESIGN AND SETTING

Retrospective cohort drawn from the Nationwide Inpatient Sample of 6. 4 million discharges from 904 representative nonfederal hospitals during 1994.

PATIENTS

All 61,223 patients in the sample whose discharge records indicated all of the following: acute respiratory distress or failure, mechanical ventilation, > or = 24 h of hospitalization, and age > or = 5 years.

RESULTS

An estimated 329,766 patients discharged from nonfederal hospitals nationwide in 1994 met study criteria for ARF. The incidence of ARF was 137.1 hospitalizations per 100,000 US residents age > or = 5 years. Incidence increased nearly exponentially each decade until age 85 years. Overall, 35.9% of patients with ARF did not survive to hospital discharge. At 31 days, hospital mortality was 31.4%. According to the proportional hazards model, significant mortality hazards included age (> or = 80 years and > or = 30 years), multiorgan system failure (MOSF), HIV, chronic liver disease, and cancer. Hospital admission for coronary artery bypass, drug overdose, or trauma other than head injury or burns was associated with a reduced mortality hazard. Interaction was present between age and MOSF, trauma, and cancer. A point system derived from the hazard model classified patients into seven groups with distinct 31-day survival probabilities ranging from 24 to 99%.

CONCLUSIONS

The incidence of ARF increases markedly with age and is especially high among persons > or = 65 years of age. Nonpulmonary hazards explain short-term (31-day) survival.

Improvements in outcomes of acute respiratory failure for patients with human immunodeficiency virus-related Pneumocystis carinii pneumonia

In the early 1990s, hospital survival among patients with human immunodeficiency virus (HIV)-related Pneumocystis carinii pneumonia (PCP) and respiratory failure was poor, approximately 20%. We examined ICU use and outcomes for patients with acute respiratory failure from PCP from 1995 to 1997. We conducted a retrospective medical record review using a random sample of 71 hospitals in seven regions of the United States. Among 1,660 patients with confirmed or presumed PCP, 155 (9%) received mechanical ventilation for respiratory failure. Factors that predicted use of mechanical ventilation, independent of severity of illness on hospital admission, included African-American ethnicity and geographic location (p </= 0.002). Hospital survival for patients receiving mechanical ventilation was 38% (95% CI 30, 46). Controlling for severity of illness, patients who were on PCP prophylaxis prior to developing PCP were less likely to survive to hospital discharge (p </= 0.02). There were no significant differences in hospital survival regardless of whether patients had received less than or more than 5 d of PCP treatment prior to respiratory failure (39 versus 29%; p = 0.5). In conclusion, from 1995 to 1997, hospital survival after PCP requiring mechanical ventilation was approximately 40%. Physicians caring for patients with severe HIV-related PCP should be aware of the improvements in outcomes for this disease before making recommendations about withholding or withdrawing ventilatory support for respiratory failure.

Clinical course, prognostic factors, and outcome prediction for HIV patients in the ICU. The PIP (Pulmonary complications, ICU support, and prognostic factors in hospitalized patients with HIV) study

STUDY OBJECTIVE

To describe the clinical course and prognostic factors in patients with HIV admitted to the ICU.

DESIGN

Prospective, observational.

SETTING

A university-affiliated medical center.

METHODS

: We included 169 consecutive ICU admissions, from April 1995 through March 1999, of 141 adults with HIV. Data collected included APACHE (acute physiology and chronic health evaluation) II score, CD4(+) lymphocyte count, serum albumin level, in-hospital mortality, and the development of organ failure, systemic inflammatory response syndrome (SIRS), and ARDS.

RESULTS

The ICU admission rate of hospitalized patients with HIV infection was 12%. The most common reason for ICU admission was respiratory failure, occurring in 65 patient admissions. Mechanical ventilation was required in 91 admissions (54%), ARDS developed in 37 admissions (22%), Pneumocystis carinii pneumonia was diagnosed in 24 admissions (14%), and SIRS developed in 126 admissions (75%). One or more organ failures developed in 131 admissions (78%). The actual and predicted mortality rates were 29.6% and 45.2%, respectively, with a standardized mortality ratio of 0.65. The most frequent immediate cause of death was bacterial infection. The CD4(+) lymphocyte count (median, 27.5 cells/microL vs 59 cells/microL; p = 0.0310) and serum albumin level (median 2.2 g/dL vs 2.6 g/dL; p = 0.0355) of nonsurvivors were lower and the APACHE II score (median, 30 vs 21; p < 0.0001) was higher, compared to those of survivors. A higher APACHE II score (odds ratio [OR], 1.11; 95% confidence interval [CI], 1.05 to 1.16) and a transfer from another hospital ward (OR, 3.03; 95% CI, 1.20 to 7.68) were independently associated with increased mortality. The median number of organ failures that developed in survivors was one, compared to four in nonsurvivors (p < 0.0001).

CONCLUSIONS

The outcome of HIV-infected patients admitted to the ICU has improved over the years. The CD4 count does not correlate with in-hospital mortality. Higher APACHE II scores and a transfer from another hospital ward are associated with a poor outcome.

A new preadmission staging system for predicting inpatient mortality from HIV-associated Pneumocystis carinii pneumonia in the early highly active antiretroviral therapy (HAART) era

A common severe complication of human immunodeficiency virus (HIV) infection has been Pneumocystis carinii pneumonia (PCP). Recently, with increasing use of PCP prophylaxis and multidrug antiretroviral therapy, the clinical manifestations of HIV infection have changed dramatically and the predictors of inpatient mortality for PCP may have also changed. We developed a new staging system for predicting inpatient mortality for patients with HIV-associated PCP admitted between 1995 and 1997. Trained abstractors performed chart reviews of 1,660 patients hospitalized with HIV-associated PCP between 1995 and 1997 at 78 hospitals in seven metropolitan areas in the United States. The overall inpatient mortality rate was 11.3%. Hierarchically optimal classification tree analysis identified an ordered five-category staging system based on three predictors: wasting, alveolar-arterial oxygen gradient (AaPO(2)), and serum albumin level. The mortality rate increased with stage: 3.7% for Stage 1, 8.5% for Stage 2, 16.1% for Stage 3, 23.3% for Stage 4, and 49.1% for Stage 5. This new staging system may be useful for severity of illness adjustment in the current era while exploring current variation in HIV-associated PCP inpatient mortality rates among hospitals and across cities.

Pneumocystis carinii pneumonia requiring intensive care management: survival and prognostic study in 110 patients with human immunodeficiency virus

OBJECTIVE

To perform a descriptive study of patients with acute respiratory failure secondary to acquired immunodeficiency syndrome-related Pneumocystis carinii pneumonia and to identify variables that are predictive of death within 3 months.

DESIGN

Case series study.

SETTING

Infectious disease intensive care unit (ICU) in a university hospital.

PATIENTS

Detailed clinical, laboratory, and ventilatory data were collected prospectively within 48 hrs of admission and during the ICU stay in 110 consecutive human immunodeficiency virus-infected patients requiring ICU management with or without mechanical ventilation for P. carinii pneumonia-related acute respiratory failure.

MEASUREMENTS AND MAIN RESULTS

Continuous positive airway pressure was used initially in 66 (60%) patients. Among the 34 patients (31%) who required mechanical ventilation, including 12 at admission and 22 after failure of continuous positive airway pressure, 76% died. The 3-month mortality rate after ICU admission was estimated at 34.6% (95% confidence interval [CI], 25%-44%). The 1-yr survival rate was estimated at 47% (95% CI, 36%-58%). With successive multiple logistic regression models analyzing the relative prognostic importance of baseline clinical and laboratory tests variables, ventilation variables, and events in the ICU, only delayed mechanical ventilation after 3 days (odd ratio [OR], 6.7; 95% CI, 1.9-23.9), duration of mechanical ventilation of > or = 5 days (OR, 2.8; 95% CI, 1.1-6.9), nosocomial infection (OR, 5.2; 95% CI, 2.1-12.9), and pneumothorax (OR, 5; 95% CI, 1.7-14.7) were predictive of death within 3 months of ICU admission. Among patients with delayed mechanical ventilation on day 3 or later and with a pneumothorax associated or not associated with a nosocomial infection, the predicted probability of 3-month death was close to 100%.

CONCLUSIONS

Our data suggest that the most significant predictive factors of death were identifiable during the course of P. carinii pneumonia-related acute respiratory failure rather than at admission and can help in bedside decisions to withdraw intensive care support in such patients.

Validation of the modified multisystem organ failure score as a predictor of mortality in patients with AIDS-related Pneumocystis carinii pneumonia and respiratory failure

STUDY OBJECTIVES

To validate a previously developed multisystem organ failure (MSOF) score with and without the addition of the lactate dehydrogenase (LDH) level as a predictor of survival to hospital discharge in patients with AIDS-related Pneumocystis carinii pneumonia (PCP) and acute respiratory failure (ARF).

DESIGN

Retrospective chart review between April 1, 1991, and September 30, 1996.

SETTING

University-affiliated tertiary care center in downtown Vancouver, British Columbia, Canada.

PATIENTS

All patients with PCP-related ARF admitted to the ICU of St. Paul's Hospital during the study period.

INTERVENTIONS

As putative prognostic instruments, data were extracted regarding the APACHE II (acute physiology and chronic health evaluation II), acute lung injury (ALI), AIDS, and modified MSOF scores, as well as LDH levels, at entry to the ICU. Patients were stratified based on an LDH level of < or > or = 2,000 U/L and this threshold was assessed in its predictability of outcome when added to each of the above scores. For APACHE II, the score was categorized in six groups and evaluated with and without inclusion of the LDH. Receiver operating characteristic curves were constructed for LDH and for each score with and without the LDH level to assess accuracy of prediction. The area under each curve was calculated and compared to estimate the statistical significance of observed differences.

MEASUREMENTS AND RESULTS

There were 40 admissions to the ICU of 38 patients with 52.5% mortality. The ALI and AIDS scores were not predictive of outcome. The modified MSOF and APACHE II scores were significant predictors of survival and the performance of both was enhanced by the addition of LDH.

CONCLUSIONS

Both the APACHE II and the modified MSOF scores were significant predictors of outcome in the patient population studied. These results validate the modified MSOF score as an effective predictor of survival to hospital discharge among patients with AIDS-related PCP who develop ARF and the performance of the score is enhanced by the addition of the LDH level.

Variations in intensive care unit utilization for patients with human immunodeficiency virus-related Pneumocystis carinii pneumonia: importance of hospital characteristics and geographic location

OBJECTIVE

To determine whether intensive care unit (ICU) use and outcomes for patients with human immunodeficiency virus (HIV)-related Pneumocystis carinii pneumonia vary by hospital characteristics and geographic location.

DESIGN

Retrospective review of the medical records of 2,174 patients with HIV-related P. carinii pneumonia.

SETTING

Random sample of 73 private, nine public, and 14 Veterans Affairs hospitals in five cities (Chicago, New York, Los Angeles, Miami, and Durham, NC).

PATIENTS

Stratified random sample of patients hospitalized with HIV-related P. carinii pneumonia from 1987 to 1990.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Among the 2,174 patients with P. carinii pneumonia, 398 (18%) patients received care in an ICU. ICU utilization varied significantly by patient and hospital characteristics, as well by as geographic location. Non-Hispanic whites, patients with Medicaid, and patients with a prior acquired immunodeficiency syndrome-defining illness were the least likely to receive care in an ICU. Patients in county- or state-owned hospitals and patients in hospitals with more P. carinii pneumonia-experience were also less likely to be cared for in an ICU. These differences in ICU utilization persisted when controlling for severity of illness, as well as other patient characteristics. Significant geographic variation in ICU utilization persisted after controlling for patient and hospital characteristics. Survival to hospital discharge after an ICU stay was significantly higher for patients without a prior acquired immunodeficiency syndrome-defining illness and for patients in hospitals with more P. carinii pneumonia experience.

CONCLUSIONS

We found significant variations in ICU utilization by hospital characteristics and geographic location that remained significant after controlling for severity of illness and patient sociodemographic characteristics. Hospital and geographic variations in ICU utilization may make it difficult to generalize ICU outcomes across different hospitals.

Care for the patient with AIDS in the ICU: outcomes and ethics

HIV infection and AIDS are common diagnoses in many intensive care units (ICUs) in the United States. Although Pneumocystis carinii currently represents only one quarter of all diagnoses for which HIV-infected persons are admitted to the ICU, it is the disease with the most clinically applicable outcome data and, therefore, is a model for ethical decision-making regarding patients with HIV infection in the ICU. Despite advances in diagnosis and treatment of HIV-related P. carinii, recent studies show that only 20% to 25% of the patients with acute respiratory failure survive to hospital discharge. Although many clinical markers correlate with survival, none of the individual markers or prediction scoring systems have the accuracy needed in clinical practice. One goal of predicting outcome in the ICU is to aid both the patient and the physician in making decisions about when to pursue aggressive therapy and when to withhold or withdraw such therapy. Because our ability to predict outcome is limited, advance directives and communication with patients and families about end-of-life medical care are of utmost importance. Even though it is not always possible for patients to predict, in advance, what they would want done in various hypothetical health care scenarios, quality communication between physicians, patients, and families with realistic discussion of outcomes and maintenance of hope and dignity can facilitate decisions about the use of intensive care for patients with AIDS.

The Relationship of the ACCP/SCCM Consensus Conference Classification of Sepsis to Mortality and Multiorgan Dysfunction among Medical ICU Patients

To determine the relation between the proposed ACCP/SCCM Consensus Conference classification of sepsis and hospital outcomes, we conducted a single-center, prospective observational study at Barnes Hospital, St. Louis, MO, an academic tertiary care hospital. A total of 324 consecutive patients admitted to the medical intensive care unit (ICU) were studied for prospective patient surveillance and data collection. The main outcome measures were the number of acquired organ system derangements and hospital mortality. Fifty-seven (17.6%) patients died during the study period. The proposed classifications of sepsis (e.g., systemic inflammatory response syndrome [SIRS], sepsis, severe sepsis, septic shock) correlated with hospital mortality ( r = 0.330; p < 0.001) and development of an Organ System Failure Index (OSFI) of 3 or greater ( r = 0.426; p < 0.001). Independent determinants of hospital mortality for this patient cohort ( p < 0.05) were development of an OSFI of 3 or greater (adjusted odds ratio [AOR], 13.9; 95% confidence interval [CI], 6.4–30.2; p < 0.001); presence of severe sepsis or septic shock (AOR, 2.6; 95% CI, 1.2–5.6; p = 0.002), and an APACHE II score ≥ of 18 or greater (AOR, 2.4; 95% CI, 1.0–5.8; p = 0.045). Intra-abdominal infection (AOR, 19.1; 95% CI, 1.6–230.1; p = 0.011), an APACHE II score ≥ of 18 or greater (AOR, 8.9; 95% CI, 4.2–18.6; p < 0.001), and presence of severe sepsis or septic shock (AOR, 2.9; 95% CI, 1.5–5.4; p = 0.001) were independently associated with development of an OSFI of 3 or greater. These data confirm that acquired multiorgan dysfunction is the most important predictor of mortality among medical ICU patients. In addition, they identify the proposed ACCP/SCCM Consensus Conference classification of sepsis as an additional independent determinant of both hospital mortality and multiorgan dysfunction.

Outcome of AIDS patients requiring mechanical ventilation predicted by recursive partitioning

Mechanical ventilatory support (VS) is often required for patients with AIDS. Patients, and/or their surrogates often ask the likely outcome of this intervention. To answer this question, we have developed a classification tree using clinical data from 71 patients with AIDS identified from the discharge abstracts of two hospitals between January 1990 and September 1994. These data were obtained at the time of hospital admission prior to any treatment and before VS was initiated. Survival was defined as discharge from the hospital that occurred in 13 of 72 admissions reviewed. A classification tree was developed by binary recursive partitioning. The output of the resulting tree was adjusted to produce a positive predictive value for death of 100% (95% confidence interval [95% CI], 94 to 100%) and a sensitivity and specificity of 98% (95% CI, 91 to 100%) and 100% (95% CI, 74 to 100%), respectively. The negative predictive value was 92% (95% CI, 64 to 100%). The tree predicted that patients with lactate dehydrogenase (LDH) levels less than 1,176 IU/L survived until hospital discharge, unless they had a positive blood culture, active tuberculosis prior to VS, a blood CD4 count less than 12 cells per cubic millimeter, or creatinine and hemoglobin values that were either above 2.4 mg/dL or less than 8.5 mg/dL, respectively. The remainder of the patients with an LDH level above 1,176 IU/L in this study died before hospital discharge. The classification tree requires prospective validation before it can be used as a predictive instrument. Nevertheless, this approach can be used to develop a concise summary of the local outcome experience of this circumstance in a manner that could be conveyed to patients and/or their surrogates.

The role of functional status in predicting inpatient mortality with AIDS: a comparison with current predictors

To assess the independent prognostic role of functional status, as reflected by a measure of an inpatient's global requirement for nursing assistance with basic activities of daily living (Global ADL), we compared Global ADL with three validated AIDS mortality predictors: the Clinical AIDS Prognostic Staging (CAPS); the Severity Classification System for AIDS Hospitalization--version 2 (SCAH-2); and CD4 cell count. Our study sample consisted of 1392 patients with AIDS and a hospital stay of 3 or more days at one of 20 hospitals in 11 U.S. cities with high AIDS incidence. Data were collected from September 1990 through December 1991. Two percent of patients refused participation, and 26% were eliminated due to incomplete data collection, leaving an analytic sample of 1003 patients. Only 30% of patients had a CD4 count measured at any time during hospitalization. Cox regression was used to measure the hazard of inpatient mortality adjusted for length of stay. Overall mortality was 12%. Mortality rates for patients in Global ADL stages I-IV were 3%, 8%, 19%, and 51%, respectively (p < 0.0001). Global ADL more effectively discriminated mortality than CAPS (p < 0.001), SCAH-2 (p < 0.001), or CD4 count (p < 0.001). Global ADL also added independent information in analyses adjusted for both CAPS and SCAH-2: a single stage increase of Global ADL demonstrated a 1.9-fold increased hazard of death (CI: 1.6, 2.3). SCAH-2, assigned at discharge, was not strongly correlated with admission predictors (Global ADL: r = 0.17; CI: 0.11, 0.23 or CAPS: r = 0.03, CI: 0.02, 0.17). We conclude that Global ADL, alone or in tandem with other severity systems, provides an excellent severity adjustment for inpatient mortality with AIDS. Finally, CD4 cell counts were not routinely available and were not as predictive as Global ADL in the patients for whom both were available.

Determinants of mortality and multiorgan dysfunction in cardiac surgery patients requiring prolonged mechanical ventilation

OBJECTIVES

To identify characteristics associated with mortality and the development of multiorgan dysfunction in patients who had undergone cardiac surgery and required prolonged mechanical ventilation, ie, > 48 h.

DESIGN

A prospective cohort study.

SETTING

Barnes Hospital, St. Louis, an academic tertiary care center.

PATIENTS OR OTHER PARTICIPANTS

A total of 107 consecutive patients undergoing cardiac surgery and requiring prolonged mechanical ventilation.

INTERVENTIONS

Prospective patients surveillance and data collection.

MAIN OUTCOME MEASURES

ICU mortality and multiorgan dysfunction.

RESULTS

Among 472 consecutive patients admitted to the cardiac surgery ICU following surgery, 107 (22.7%) required prolonged mechanical ventilation. Twenty-one of these patients (19.6%) died during their hospitalization. In a logistic-regression analysis, the development of an organ system failure index (OSFI) of 3 or greater was the only characteristic independently associated with ICU mortality (p < 0.001). The occurrence of an antibiotic-resistant infection (adjusted odds ratio [AOR] = 6.1, 95% confidence interval [CI] = 2.5 to 14.6 p = 0.006), an aortic cross-clamp time equal to or greater than 1.25 h (AOR = 3.9, CI = 2.3 to 6.8, p = 0.016), the development of ventilator-associated pneumonia (AOR = 3.6, CI = 2.4 to 5.3, p < 0.001), and an APACHE III score equal to or greater than 30 (AOR = 3.1, CI = 1.8 to 5.3, p = 0.036) were independently associated with the development of an OSFI of 3 or greater.

CONCLUSIONS

These data confirm that acquired multiorgan dysfunction is the best predictor of mortality in patients requiring prolonged mechanical ventilation following cardiac surgery. Additionally, they identify potential determinants of multiorgan dysfunction and suggest possible interventions for its reduction in this patient population.

Decreasing frequency but worsening mortality of acute respiratory failure secondary to AIDS-related Pneumocystis carinii pneumonia

OBJECTIVE

To describe changes in incidence and outcome of acute respiratory failure (ARF) due to AIDS-related Pneumocystis carinii pneumonia (PCP) at a tertiary care center over the 4-year period starting April 1, 1987 with reference to previously reported data from the preceding 6 years.

METHODS

All patients admitted to St. Paul's hospital with a diagnosis of AIDS-related PCP during the study period were reviewed with regard to diagnostic, clinical, therapeutic, and outcome variables.

RESULTS

A total of 456 episodes of PCP were diagnosed during the study period. These were compared against 127 cases diagnosed between 1981 and 1987. The frequency of hospitalization for PCP decreased to 78% in 1987 to 1991 from 100% in 1981 to 1987 (p < or = 0.001). A similar decreasing trend was observed with regard to the incidence of PCP-related ARF that declined from 21% in 1981 to 1987 to 9% in 1987 to 1991 (p = 0.009). Despite this, overall PCP-related mortality remained stable at 12% in 1981 to 1987 and 9% in 1987 to 1991 (p = 0.26). The proportion of patients with PCP-related ARF who received mechanical ventilation decreased from 89% in 1981 to 1987 to 64% in 1987 to 1991 (p < 0.001). Despite this, the case fatality rate among mechanically ventilated patients increased from 50% in 1981 to 1987 to 89% in 1987 to 1991 (p = 0.003). These changes were associated with a significant change in the pattern of use of corticosteroids as adjunctive therapy for AIDS-related PCP. In 1985 to 1986, nearly 100% of patients admitted to the ICU received corticosteroids only after admission to the ICU, following the development of ARF. In contrast, in 1989 to 1990, 50% of patients were admitted to the ICU already receiving systemic corticosteroids. The rise in the proportion of patients receiving corticosteroids prior to ICU admission between these two intervals was statistically significant (p = 0.017).

CONCLUSION

Our data show a decreasing frequency but a worsening mortality of ARF secondary to AIDS-related PCP. We conclude that ARF secondary to AIDS-related PCP developing despite maximal therapy, including adjunctive corticosteroids, carries a dismal prognosis.

A simple multiple system organ failure scoring system predicts mortality of patients who have sepsis syndrome

A simple multiple system organ failure (MSOF) score may predict mortality of patients who have sepsis syndrome. Using an MSOF scoring system, we prospectively determined the presence or absence of respiratory, cardiovascular, renal, hepatic, gastrointestinal, hematologic, and neurologic organ failure on day 1 of sepsis syndrome in 154 consecutive patients who had sepsis syndrome in the ICU of a tertiary care, teaching hospital. We used 30-day hospital mortality as the primary outcome variable. Overall 30-day mortality was 34 percent. There was a strong linear association between number of organ system failures and 30-day mortality (p < 0.0001). Mortality was 20 percent in patients who had less than 3 organ system failures (n = 111) and 70 percent in patients who had 3 or more organ system failures (n = 43). Survival was assessed using the Cox proportional hazards model and was found to be significantly different (p < 0.01) between the two groups defined by the aforementioned dichotomy after adjustment for age and sex using time to death as the primary outcome. The increase in relative risk of death associated with 3 or more organ system failures was 2.77 (95 percent confidence interval, 2.74 to 2.83). Using logistic regression, the adjusted odds ratios (OR) for covariates most predictive of mortality were hematologic (OR = 6.2), neurologic (OR = 4.4), hepatic (OR = 3.4), cardiovascular (OR = 2.6), and age (1.05 per year). The logistic model using the seven organ system failures and age as covariates accurately predicted outcome 75 percent of the time with a sensitivity of 51 percent and specificity of 87 percent. In conclusion, a simple scoring system tabulating the number of organ system failures present on day 1 of sepsis syndrome predicts the mortality of patients who have sepsis syndrome with reasonable accuracy.