Treatment of cefuroxime-induced neurotoxicity with continuous venovenous haemofiltration.

A 61-year-old woman with decreased consciousness, myoclonus, tremors, nystagmus and bradypnoea, due to cefuroxime-induced neurotoxicity, was admitted to the intensive care unit. Continuous venovenous haemofiltration (CVVH) rapidly reduced plasma cefuroxime concentrations and improved neurological manifestations within the next few hours. Retrospective pharmacokinetic assessment showed a total cefuroxime clearance of 166 ml/min during the CVVH.

Simultaneous determination of levomepromazine, midazolam and their major metabolites in human plasma by reversed-phase liquid chromatography.

A sensitive and reliable high-performance liquid chromatographic (HPLC) assay is a prerequisite for pharmacokinetic analysis of continuous infusion of levomepromazine adjuvant to midazolam. We developed such a method to determine the levels of levomepromazine, midazolam and their major metabolites (levomepromazinesulfoxide, desmethyl-, didesmethyllevomepromazine, O-desmethyllevomepromazine and alpha-hydroxy-midazolam) simultaneously. Desmethylclomipramine was used as an internal standard (I.S.). The lower limit of quantification of this assay was set for levomepromazine 4.1 microg/l, levomepromazinesulfoxide 4.9 microg/l, O-desmethyllevomepromazine 18.4 microg/l, alpha-hydroxymidazolam 26.6 microg/l, midazolam 23.4 microg/l, didesmethyllevomepromazine 15.8 microg/l, and desmethyllevomepromazine 6.6 microg/l. The between- and within day assay variations were commonly below 5%. The recovery in human plasma for the different analytes varied between 85 and 11%. The accuracy of this assay varied between 95 and 105% for the different concentrations. The linearity of this assay was set between 25 and 800 microg/l (r(2)>0.999 of the regression line). The first results of pharmacokinetic analysis of midazolam indicated that half-life varied between 1.1 and 1.9 h. Pharmacokinetic analysis using a one-compartment model of levomepromazine revealed that the apparent volume of distribution was 4.1+/-2.4 l per kg lean body mass and the metabolic clearance was 309+/-225 l per hour per 70 kg. This assay proved to be robust and reproducible. It can reliably be used for further study of the pharmacokinetics of continuous infusion of levomepromazine.

Durable HIV-1 suppression with indinavir after failing lamivudine-containing double nucleoside therapy: a randomized controlled trial.

OBJECTIVE: To assess the durability of the antiretroviral effect in plasma and cerebrospinal fluid (CSF) of antiviral therapy intensification, produced by the addition of indinavir from week 12 onwards to the original regimen of zidovudine/lamivudine or stavudine/lamivudine, after 72 weeks of follow-up using an ultrasensitive HIV-1 RNA assay. To assess CSF concentrations of indinavir at week 48. DESIGN: In a prospectively, randomized, open, single-centre study, antiretroviral-naive patients (CD4 cell count > or =200 cells/microl and a plasma HIV-1 RNA level 10,000 copies/ml) were assigned to a combination of zidovudine/lamivudine or stavudine/lamivudine. Indinavir could be added to the double nucleoside analogue regimen from week 12 or thereafter in case the plasma HIV RNA level was insufficiently suppressed (>500 copies/ml). RESULTS: Forty-seven patients were enrolled (23 stavudine/lamivudine and 24 zidovudine/lamivudine), of whom 33 completed a follow-up of 72 weeks. Indinavir was added in 89% (42/47) of the patients. Only one discontinuation occurred due to virological failure. At week 72, the median plasma HIV-1 RNA levels in the zidovudine/lamivudine group had decreased from 4.80 log10 copies/ml to <500 copies/ml in 100% of patients and <50 copies/ml in 86.6% of the patients. In the stavudine/lamivudine group the plasma HIV-1 RNA decreased from 4.98 log10 copies/ml at baseline to <500 copies/ml in 100% of patients and <50 copies/ml in 66.7% of the patients. On an intent-to-treat basis these figures were 54.2 and 52.2% for zidovudine/lamivudine and stavudine/lamivudine, respectively, for the 50 copies/ml assay. The median CD4 cell count increased from 315 cells/microl, with 150 cells/microl in the zidovudine/lamivudine arm, and from 290 cells/microl, with 310 cells/microl in the stavudine/lamivudine arm (P=0.0001). However, the percentage of CD4 cells did not differ in each group. In the zidovudine/lamivudine group 9/10 and 5/5, and in the stavudine/lamivudine group 11/11 and 6/6 had a CSF HIV-1 RNA level <50 copies/ml at week 12 and 48, respectively. The CSF indinavir concentration ranged from 50 to 170 ng/ml. CONCLUSION: The long-term HIV-1 suppression observed in this study is remarkable, as adding a single antiretroviral agent to a failing regimen goes against current notions of adequate therapy.

Clinical features and risk factors of lactic acidosis following long-term antiretroviral therapy: 4 fatal cases.

Our objective was to describe clinical features and predisposing factors attributed to lactic acidosis in 4 HIV-infected patients on long-term nucleoside reverse transcriptase inhibitor (NRTI) therapy. All patients had received at least 6-20 months of NRTI-containing antiretroviral therapy: all used stavudine (d4T), in one combined with lamivudine (3TC), in the other 3 with didanosine (ddI); in one hydroxyurea was added. In all, the initial symptoms were gastrointestinal (nausea and vomiting), followed by tachypnoea preceding the lactic acidosis; death followed 6-22 days after admission (liver failure and uncontrollable arrhythmias). Treatment with riboflavin was unsuccessful in one patient. The only definite risk factor in all cases was NRTI-induced mitochondrial toxicity; one patient was concomitantly treated for Kaposi's sarcoma (with bleomycin and vinblastine) and one just recovered from pneumococcal sepsis. None of the patients had a history of chronic hepatitis B virus (HBV) or hepatitis C virus (HCV) infection. In all patients, some sort of toxicity to other previously used NRTIs had occurred earlier. Lactic acidosis occurred after months of NRTI therapy in patients who had already suffered other forms of NRTI toxicity. Concomitant diseases or comedication might have aggravated the mitochondrial toxicity of the NRTIs. Screening methods to detect mitochondrial toxicity are necessary, since lactic acidosis occurs rather unexpectedly, with a rapid, fatal course.

The steady-state pharmacokinetics of nevirapine during once daily and twice daily dosing in HIV-1-infected individuals.

OBJECTIVE: To investigate and to compare the steady-state plasma pharmacokinetics of nevirapine in a dosing regimen of 400 mg once daily versus 200 mg twice daily in HIV-1-infected individuals. DESIGN: Open-label, randomized, cross-over study. METHODS: Twenty HIV-1-infected individuals who already used nevirapine as part of their antiretroviral regimen were randomized to continue their current regimen (200 mg twice daily) or to switch to the alternate regimen (400 mg once daily). The steady-state plasma pharmacokinetics of nevirapine were assessed after 2 weeks during a 24-h period. Subsequently, patients were switched to the alternate regimen and the pharmacokinetics of nevirapine were assessed again after 2 weeks. Non-compartmental methods were used to calculate the area under the plasma concentration versus time curve (AUC[24h]), and the maximal (Cmax) and minimal plasma concentration (Cmin), the time to Cmax (t(max)), the plasma elimination half-life (t1/2), the apparent oral clearance (Cl/F) and the apparent volume of distribution (V/F). Differences in these pharmacokinetic parameters for the two dosing regimens were tested using ANOVA. RESULTS: The exposure to nevirapine, as measured by the AUC[24h], was not significantly different between the 400 mg once daily and 200 mg twice daily dosing regimen (P = 0.60). Furthermore, the values for t(max), t1/2 Cl/F and V/F were not significantly different between the two dosing regimens (P > or = 0.08). However, Cmax and Cmin were higher and lower, respectively, when nevirapine was used in the once daily regimen as compared with the twice daily regimen. The median values for Cmax and Cmin as measured for the once daily and twice daily regimens were 6.69 and 5.74 microg/ml, respectively (P = 0.03), and 2.88 and 3.73 microg/ml, respectively (P < 0.01). CONCLUSION: These data show that the daily exposure to nevirapine, as measured by the plasma AUC[24h], is not different between a 400 mg once daily and a 200 mg twice daily dosing regimen. However, Cmax and Cmin are higher and lower, respectively, for the once daily regimen as compared with the twice daily regimen. The clinical implications of these differences remain to be established.

Cerebrospinal fluid beta2-microglobulin, monocyte chemotactic protein-1, and soluble tumour necrosis factor alpha receptors before and after treatment with lamivudine plus zidovudine or stavudine.

CSF levels of beta2-microglobulin (b2m), monocyte chemotactic protein-1 (MCP-1), soluble tumor necrosis factor receptors (sTNFRs), and HIV-1 RNA were determined in 16 neurologically asymptomatic HIV-1 infected patients before and 12 weeks after treatment with lamivudine plus zidovudine or stavudine. b2m levels were significantly higher in patients (1.7 mg/l) compared with controls (0.8 mg/l) (P < 0.001), and decreased to 1.1 mg/l during treatment (P = 0.001). MCP-1 levels were low, and did not change during treatment. Levels of sTNFR type I were elevated in patients (0.92 ng/ml) compared to controls (0.30 ng/ml) (P = 0.03), but did not change during treatment. Levels of sTNFR type II were below the limit of detection in most patients and controls. In conclusion, CSF levels of b2m and HIV-I RNA, but not sTNFRs or MCP-1, are candidate surrogate markers of treatment efficacy in early CNS infection.

Immune reconstitution after 2 years of successful potent antiretroviral therapy in previously untreated human immunodeficiency virus type 1-infected adults.

Today's antiretroviral combination regimens can induce significant and sustained decreases in human immunodeficiency virus (HIV)-RNA levels, allowing the immune system to recover. To what extent immune reconstitution is possible and what factors determine the outcome have thus far not been resolved. We studied 19 subjects, treated for 2 years with protease inhibitor-containing triple therapy, who had a strong suppression of HIV-RNA levels. CD4+ T-cell numbers increased from medians of 170 to 420x106 cells/L, but in a number of subjects T-cell numbers did not further increase after week 72, without having reached normal values. Long-term CD4+ T-cell change was mainly caused by a slow but continuous increase in naive CD4+ T cells (CD45RA+CD62L+) and was predicted by the baseline number of these cells. Our data indicate that long-term immunological recovery is gradual, even during strong suppression of viral replication, not always complete, and dependent on the preexisting level of naive CD4+ T cells.

Functional T cell reconstitution and human immunodeficiency virus-1-specific cell-mediated immunity during highly active antiretroviral therapy.

Lymphoproliferative responses (LPRs) to recall antigens (Ags) and human immunodeficiency virus type 1 (HIV-1) Gag and frequencies of circulating HIV-1-specific cytotoxic T lymphocyte precursors (CTLps) were measured in 12 patients undergoing highly active antiretroviral therapy (HAART) after long-standing HIV-1 infection. LPRs to at least 1 recall Ag became detectable or increased in all patients during HAART. No significant LPRs to Gag-p24 were observed, whereas 4 of 8 patients tested presented with Gag-p17-specific LPRs. HIV-1-specific CTLp frequencies became measurable or increased early during therapy in 6 of 10 patients tested and were maintained or decreased thereafter. Increasing HIV-1-specific CTLp frequencies were seen only in association with partial HAART failure in 1 patient. In conclusion, restoration of CD4+ T lymphocyte responsiveness to recall Ags is achieved during HAART. The data provide evidence for limited HIV-1-specific CD4+ memory T cells during advanced HIV-1 infection and suggest that both CD4+ and CD8+ HIV-1-specific T cells are poorly stimulated when viral load is suppressed.

Immunopathology as a result of highly active antiretroviral therapy in HIV-1-infected patients.

OBJECTIVE: Unusual clinical inflammatory syndromes associated with underlying previously unrecognized opportunistic infections are increasingly being noted shortly after starting highly active antiretroviral therapy (HAART). This study examined the possible relationship between such unexpected disease manifestations and in vitro parameters of microbial antigen-specific immune reactivity in patients infected with HIV-1 who had a Mycobacterium avium intracellulare or Mycobacterium xenopi infection. DESIGN: In vitro T-cell proliferation experiments were performed after specific stimulation of a patient's peripheral blood mononuclear cells (PBMC) with M. avium and M. xenopi antigen and non-specific stimulation with phytohaemagglutinin (PHA). The results were compared with appropriate controls. PATIENTS: Five patients who presented with unusual clinical syndromes associated with M. avium or M. xenopi infection within weeks of experiencing large rises in CD4+ cell counts following the initiation of antiretroviral therapy. RESULTS: In all patients except one, mycobacteria-specific lymphoproliferative responses rose significantly following HAART; this was temporally associated with elevations in CD4+ cell counts and the occurrence of clinical disease. The patient with M. xenopi infection appeared to clear his infection subsequently without antimycobacterial therapy. In three of the four patients with M. avium infection, antimycobacterial treatment could be stopped without recurrence of infection. CONCLUSION: Our findings support the hypothesis that HAART may lead to clinically relevant inflammation as a result of restoration of specific immune reactivity against microbial pathogens that are subclinically present at the time treatment is initiated. Continuation of HAART may subsequently result in protective immunity and clearance of infection.

An open randomized controlled trial of zidovudine plus lamivudine versus stavudine plus lamivudine.

OBJECTIVE: To compare the antiretroviral effect and safety of zidovudine (ZDV)-lamivudine (3TC) with that of stavudine (d4T)-3TC. METHODS: In an open randomized controlled trial antiretroviral therapy-naive patients who had CD4+ counts > or = 200 x 10(6)/l and plasma HIV RNA load > or = 10000 copies/ml were randomized to receive ZDV-3TC (200 mg three times daily and 150 mg twice daily, respectively) or d4T-3TC (40 mg and 150 mg, both twice daily). If the plasma HIV RNA level at week 8 or thereafter was > 500 copes/ml, indinavir was added at the next scheduled visit. Genotypic resistance analysis of the reverse transcriptase gene was performed at week 0 and 12. Results over 24 weeks were reported. RESULTS: Forty-seven patients were treated (24 took ZDV-3TC; 23 took d4T-3TC). Plasma HIV RNA levels decreased from median 4.80 to 3.15 log10 copies/ml (ZDV-3TC, P < 0.0001) and from 4.98 to 3.03 log10 copies/ml (d4T-3TC, P < 0.0001) after 12 weeks of treatment. Indinavir was added at week 12 in 11 out of 21 patients with ZDV-3TC and in 10 out of 22 patients with d4T-3TC. Median virus load at week 24 was 2.41 log10 and 2.29 log10 copies/ml (P=0.14), respectively. Seventy-five per cent (15 out of 20; ZDV-3TC) and 95% (18 out of 19; d4T-3TC) of patients had a virus load of < 500 copies/ml. Genomic evidence for 3TC resistance was found in all patients tested (11/11 ZDV-3TC and 12/12 d4T-3TC). At week 12, CD4 cell counts had increased with a median of 110 x 10(6)/l in the ZDV-3TC group (baseline, 315 x 10(6)/l) and a median of 115 x 10(6)/l in the d4T-3TC group (baseline 290 x 10(6)/l). At week 24, the median increases were 90 and 120 x 10(6)/l, respectively. Overall the increase of CD4+ cells was higher in the d4T-3TC group (P=0.02). CONCLUSION: d4T-3TC is at least as effective as ZDV-3TC, but 3TC resistance emerged in all patients investigated. The virological response of the dual nucleoside combination is of short duration. However, after addition of indinavir the virus load could be reduced to < 500 copies/ml in the majority of patients. The increase in CD4+ cell count was significantly greater in the d4T-3TC group. To prevent 3TC resistance, the drug should not be used in regimens containing only two nucleosides, irrespective the virus load at baseline.

Diagnosis of Mycobacterium microti infections among humans by using novel genetic markers.

As a result of DNA typing of Mycobacterium microti isolates from animals in the United Kingdom and The Netherlands, we diagnosed four human M. microti infections. These are the first M. microti infections among humans to be reported. Three of the patients were immunocompromised and suffered from generalized forms of tuberculosis. The fourth patient was a 34-year-old immunocompetent male with a persistent cough and undefined X-ray abnormalities. Two of the M. microti infections were recognized by their IS6110 restriction fragment length polymorphism (RFLP) patterns, which showed a high degree of similarity with those of M. microti strains isolated from a pig and a ferret in The Netherlands. The two other human M. microti infections were recognized by using the recently developed DNA fingerprinting method, "spoligotyping," directly on clinical material. All M. microti isolates from the United Kingdom and The Netherlands were found to contain an exceptionally short genomic direct repeat region, resulting in identical two-spacer sequence reactions in spoligotyping. In contrast, the highly similar IS6110 RFLP patterns of the vole strains from the United Kingdom differed considerably from the RFLPs of all M. microti strains isolated in The Netherlands, suggesting that geographic isolation led to divergent strains in the United Kingdom and on the continent.

Decrease of HIV-1 RNA levels in lymphoid tissue and peripheral blood during treatment with ritonavir, lamivudine and zidovudine. Ritonavir/3TC/ZDV Study Group.

OBJECTIVES: Triple combination treatment of HIV-1 infection using two reverse transcriptase inhibitors and a protease inhibitor can result in significant and sustained decreases in the quantity of viral RNA in peripheral blood. Lymphoid tissue, however, constitutes the major reservoir of HIV in infected patients. Study of the viral burden in these tissues has provided additional insight in the efficacy of antiretroviral treatment. DESIGN: Patients were randomized into two groups in order to study differences in the development of resistance to reverse transcriptase inhibitors. Group I started treatment with all three drugs simultaneously. Group II started with ritonavir monotherapy, aiming at initial reduction in virus production before the addition of lamivudine and zidovudine 3 weeks later. METHODS: Changes in the amount of HIV in plasma and tonsillar lymphoid tissue during 24 weeks of treatment with ritonavir, lamivudine and zidovudine were studied by reverse transcriptase polymerase chain reaction. RESULTS: Thirty-three antiretroviral-naive HIV-infected patients were included for analysis. After 24 weeks, median CD4+ cell count increased by 152 x 10(6)/l and median plasma viral RNA levels decreased by at least 2.87 log10 copies/ml. In 88% of the patients remaining on treatment, plasma RNA levels were below the quantification limit of the assay used (mean, 2.4 log10 copies/ml). The lymphoid tissue viral burden, ranging from 9.16 to 8.52 log10 copies/g at baseline, was markedly reduced with at least 2.1 log10 copies/g by week 24 in the five patients analysed. Eight patients (24%) withdrew because of side-effects. In one patient in group II, ritonavir and lamivudine resistance-associated mutations developed. CONCLUSIONS: Treatment with this triple antiretroviral drug combination produced a durable and strong decrease of HIV-1 RNA burden in both plasma and lymphoid tissue.

Improvement of chronic diarrhoea in patients with advanced HIV-1 infection during potent antiretroviral therapy.

BACKGROUND: A substantial number of patients with advanced HIV infection suffer from intractable diarrhoea. The aim of this study was to evaluate whether potent antiretroviral therapy could alleviate such diarrhoea. METHODS: In an open randomized study the effect of the HIV protease inhibitor indinavir in combination with nucleoside analogue reverse transcriptase inhibitors on chronic HIV-related diarrhoea was investigated in 14 late-stage (CD4+ lymphocyte count < or = 50 x 10(6) cells/l) HIV-infected patients. Data concerning stool frequency, stool consistency and antidiarrhoeal drug use were collected in daily diaries over a 24-week period. Endpoints of the study were reduction of stool frequency, improvement of stool consistency, weight gain, and in case of diarrhoea due to Enterocytozoon bieneusi or Cryptosporidium sp. disappearance of these parasites from stool. RESULTS: Thirteen patients started the study drug indinavir. One patient died after 1 week and one patient withdrew prematurely after 18 weeks. Median stool frequency declined from 5.8 daily at baseline to 2.3 daily after 24 weeks (P=0.04). Stool consistency improved considerably over the study period: before treatment 56% of stools were watery and 0% were formed; at week 24 these figures were 0 and 35%, respectively. Body weight increased significantly with a median increment of 6.6 kg at week 24 (P=0.0006). In two out of six patients with microsporidiosis and both patients with cryptosporidiosis, stools were free of parasites at week 24. Five out of six patients who used non-specific antidiarrhoeal medication on a regular basis prior to the study had ceased to do so at the end. CONCLUSION: The use of potent antiretroviral therapy in patients with advanced HIV infection can improve chronic HIV-related diarrhoea and in some cases lead to disappearance of E. bieneusi and Cryptosporidium sp. from the stools.

Cerebrospinal-fluid HIV-1 RNA and drug concentrations after treatment with lamivudine plus zidovudine or stavudine.

BACKGROUND: Treatment and prevention of HIV-1-related central-nervous-system disease may be dependent on penetration of antiretroviral drugs into the central nervous system. Few data are available about cerebrospinal-fluid penetration and concomitant changes of HIV-1-RNA concentrations during treatment with antiretroviral agents. We investigated these effects in HIV-1-infected people. METHODS: 28 antiretroviral-naive individuals with CD4 cell counts of 200/microL or more and plasma HIV-1-RNA concentrations of 10,000 or more copies/mL who were free of neurological symptoms were randomly assigned lamivudine plus either stavudine (n = 17) or zidovudine (n = 11). We did lumbar punctures on 28 individuals before and 22 individuals after 12 weeks of treatment to assess HIV-1-RNA and drug concentrations in cerebrospinal fluid. FINDINGS: All 28 individuals had detectable HIV-1-RNA concentrations in the cerebrospinal fluid (median 4.64 log10 copies/mL and 4.20 log10 copies/mL in the lamivudine plus zidovudine and lamivudine plus stavudine groups, respectively). There was no correlation between plasma and cerebrospinal-fluid HIV-1-RNA concentrations (r = 0.18, p = 0.35). After 12 weeks of treatment none of the individuals had detectable HIV-1-RNA concentrations in the cerebrospinal fluid. The highest drug concentration in the cerebrospinal fluid was for lamivudine followed by stavudine and zidovudine. Concentrations were consistent over time, unlike plasma concentrations. Therefore, we found time-dependent cerebrospinal-fluid to plasma drug-penetration ratios, which were highest for zidovudine followed by stavudine and lamivudine. INTERPRETATION: The two drug combinations were equally effective in the decrease of cerebrospinal fluid HIV-1-RNA concentrations. All drugs penetrated the cerebrospinal fluid. Antiretroviral drugs other than zidovudine might be useful in the prevention of AIDS dementia complex.

Overshoot of HIV-1 viraemia after early discontinuation of antiretroviral treatment.

OBJECTIVE: To determine whether, as predicted by predator-prey dynamics, early withdrawal of antiretroviral therapy, i.e. when the number of CD4+ lymphocytes is still elevated, results in an overshoot of HIV-1 viraemia due to infection of increased numbers of available target cells at that time. DESIGN AND METHODS: Five HIV-1-infected individuals were identified who discontinued antiretroviral therapy for various reasons after 8-19 days, and from whom stored serum samples obtained before, during, and shortly after treatment were available for measurement of HIV-1 RNA load. A mathematical model was designed to assess whether increased target cell availability could quantitatively explain the clinical observations. RESULTS: After therapy withdrawal, increases in the HIV-1 RNA load to levels exceeding pretreatment values by log10 0.6-1.5 copies/ml were observed after 2-17 days in all four of the individuals who had treatment-induced increases in CD4+ cell counts at the time of therapy withdrawal. Increases in viraemia were maximal within a few days, and subsequently seemed to wane until the pretreatment equilibrium between virus and its target cells was attained. Mathematical modelling confirms that these transient increases in viraemia can be explained by increased availability of target cells at the time of therapy withdrawal. CONCLUSIONS: Transient rises in HIV-1 viraemia do occur following early therapy withdrawal. These rises especially warrant consideration in short-term antiretroviral regimens for prevention of mother-to-child transmission, as are being studied in developing countries, since they could result in an increased transmission risk during the post-partum period through breast-feeding. This possibility needs to be investigated urgently.

Candidemia in intensive care unit patients: risk factors for mortality.

Aim of this study was to evaluate whether risk factors which predict the development of candidemia may also predict death in ICU patients with candidemia. During an 8-year-period all ICU patients whose blood cultures yielded Candida species (n = 40) were retrospectively evaluated in a case-control fashion. The average incidence of Candida bloodstream infections was 5.5 per 10,000 patient days, ranging from 2.4 in 1990 to 7.4 in 1994. C. albicans was the most common pathogen in candidemic patients, but the proportion of non-C. albicans strains showed an increasing trend during 1989-1993, with a major shift towards non-C. albicans species in 1994. The overall mortality of patients with candidemia was 58%. Mortality was highest in the group of patients with multi-organ dysfunction syndrome, especially among those in need of hemodialysis. Risk factors for the development of candidemia, such as age, malignancy, steroid use, i.v. catheterization, and the use of broad-spectrum antibiotics were not correlated with mortality in the ICU patients studied.