The value of protective isolation in preventing nosocomial infections in high risk patients

Impact of intensified contact precautions while treating hematopoietic stem cell transplantation recipients during aplasia

BACKGROUND

Bacterial infections are a major complication for patients undergoing allogeneic hematopoietic stem cell transplantation (HCT). Therefore, protective isolation is considered crucial to prevent nosocomial infections in this population. Here, the impact of intensified contact precautions on environmental contamination and the occurrence of bloodstream infections (BSI) in patients on a HCT unit were compared between two contact precaution measures.

METHODS

A 2-year retrospective observational study was performed. In the first year, strict contact precaution measures were applied (i.e., protective isolation, the use of sterile personal protective equipment (PPE) by healthcare workers and visitors and sterilization of linen and objects that entered the patient's room). After one year, contact precautions were reduced (i.e., no use of sterile PPE, no sterilization of linen and objects that entered the patient's room). Environmental contamination in randomly selected patient rooms was monitored by sampling six standardized environmental sites in the respective patient treatment units. In a before-and-after study, the number of BSI episodes of those patients, who were accommodated in the monitored rooms was compared.

RESULTS

In total, 181 treatment units were monitored. No significant difference in the contamination of anterooms and patient's rooms between both groups was found. A total of 168 patients were followed for the occurrence of BSI during the entire study period (before: 84 patients, after: 84 patients). The total count of patients with BSI episodes showed a higher incidence in the period with reduced contact precautions (30/84 vs. 17/84, p = 0.039). The cause of this increasing number of BSI can be traced back to BSI episodes with common commensal bacteria (17/84 vs. 5/84, p = 0.011).

CONCLUSIONS

The implementation of maximal barrier measures did not reduce the bacterial contamination of the patients' environment. The impact on the patients' outcomes remain controversial. Further research is needed to investigate the impact of infection prevention measures on the clinical outcome of patients undergoing HCT.

Empiric Therapy of Infections in Hematologic Malignancies: A Prospective, Randomized Trial

Patients with hematologic malignancies were randomly assigned to receive cefuroxime (group A) or tobramycin plus ampicillin (Group B) during 86 febrile episodes. In both regimens carbenicillin was added during neutropenia (71% of all episodes: groups C and D). The most common type of infection was pneumonia (48% alone; 72% with other sites involved), which accounted for a high fatality rate (15%); the highest rate occurred during septicemia with pneumonia (50%). The overall response rate to initial therapy was 63% without significant differences among the four regimens. The worst prognosis was observed in neutropenic patients without granulocyte recovery. When initial and cross-over trials were combined, there were favorable outcomes in 90% of all cases. Cefuroxime alone seems to be as effective as tobramycin plus ampicillin in the treatment of infections in hematologic malignancies. No side effects could be attributed to the cefuroxime-containing regimens.

Footwear Exchange Has No Influence on the Incidence of Febrile Neutropenia in Patients Undergoing Chemotherapy for Hematologic Malignancies

Objective:

To determine whether footwear exchange affects the incidence of febrile neutropenia among patients undergoing chemotherapy for hematologic malignancies.

Design:

Open trial with historical comparison.

Setting:

The 12-bed high-efficiency particulate air-fil-tered hematology unit at Osaka University Hospital, Suita, Japan.

Patients:

Those with hematologic malignancies who underwent chemotherapy from January 1997 through January 2003. Footwear exchange was discontinued in January 2000.

Methods:

The surveillance system was based on the National Nosocomial Infections Surveillance System of the Centers for Disease Control and Prevention. Rates of febrile neutropenia were calculated for neutropenic patient-days (ie, days with neutropenia < 500/μL).

Results:

From January 1997 through December 1999 and from February 2000 through January 2003, 58 and 54 patients endured 237 and 184 neutropenic periods following chemotherapy, and their total neutropenic days were 3,123 and 2,503, respectively. They showed episodes of febrile neutropenia 89 and 68 times, respectively. Infection rates were 28.5 and 27.2 per 1,000 neutropenic patient-days (P = .83), respectively.

Conclusion:

The incidence of febrile neutropenia was not affected by footwear exchange. In hematology units, changing shoes does not appear to affect the rate of infections during neutropenic periods.

Evidence-Based Nursing Practice to Prevent Infection in Hospitalized Neutropenic Patients With Cancer

PURPOSE/OBJECTIVES

To review studies that have assessed the effectiveness of selected nursing interventions used in hospitals to prevent healthcare-associated infections in neutropenic patients with cancer.

DATA SOURCES

Literature review of low microbial diets, protective clothing and environments, personal hygiene, and oral care in English-language articles from PubMed; the Cumulative Index of Nursing and Allied Health Literature; the National Guideline Clearinghouse, 1980-June 2003; and Cochrane Database of Systematic Reviews.

DATA SYNTHESIS

Few studies have demonstrated the effectiveness of low microbial food and water and protective environments and clothing in reducing infections in neutropenic patients with cancer, and hospitals vary in these practices. Skin antisepsis reduces microbial counts, but data regarding the effect on infections are lacking. Many studies were characterized by insufficient sample sizes or use of multiple interventions.

CONCLUSIONS

Major gaps exist in empirical evidence regarding which nursing interventions might be helpful in preventing or controlling healthcare-associated infections in neutropenic patients.

IMPLICATIONS FOR NURSING

Although the evidence base for clinical practices such as a low microbial diet, protective environments and clothing, and special skin antisepsis regimens is weak, some of these practices seem prudent and reasonable. Until further evidence is available, clinicians can use consensus guidelines and should assist in identifying clinical practices that require additional research. Ultimately, interventions with little or no demonstrated efficacy should be examined systematically or abandoned. Additional studies of sufficient sample size regarding nursing practices such as the role of protective environments, room placement, antiseptic bathing, and prevention and treatment of oral complications are indicated. Because of difficulties in randomization and risk stratification, rigorous observational studies often may be an acceptable alternative to clinical trials.

The Effect of the Neutropenic Diet in the Outpatient Setting: A Pilot Study

PURPOSE/OBJECTIVES

To determine whether use of the neutropenic diet in the outpatient setting decreases the number of febrile admissions and positive blood cultures associated with chemotherapy-induced neutropenia.

DESIGN

Descriptive pilot study.

SETTING

Outpatient chemotherapy unit of a medical center in a major metropolitan area.

SAMPLE

Convenience sample of 28 patients aged 33-67 years beginning treatment with 1 of 13 chemotherapy regimens.

METHODS

Twelve-week prospective study beginning on day 1 of chemotherapy cycle 1. Patients received instructions regarding the neutropenic diet before starting chemotherapy. Adherence assessment telephone calls were made at weeks 6 and 12. Hospital admission charts were reviewed at study completion.

MAIN RESEARCH VARIABLES

Adherence with neutropenic diet, number of febrile admissions, and number of positive blood cultures.

FINDINGS

Sixteen patients were compliant with the neutropenic diet, four of which were admitted for neutropenia with gram-negative rods. No significant differences were found in the rates of febrile admissions or positive blood cultures between compliant and noncompliant patients.

CONCLUSIONS

Clinical significance in this pilot study is related to the time required for diet education, content of diet education regarding food restrictions, and difficulty adhering to diet requirements given the multitude of side effects (e.g., nausea, vomiting, mouth sores, diarrhea) of chemotherapy.

IMPLICATIONS FOR NURSING

No clear evidence exists that the neutropenic diet makes a difference in overall rates of infection. Nursing research to compare the neutropenic diet with a less restrictive food safety education-focused diet is needed to guide clinical practice.

Dietary recommendations for neutropenic patients

OBJECTIVES

To describe the development of an evidence-based practice project that will evaluate less restrictive dietary practices and focus on hand washing in adult patients with chemotherapy-induced neutropenia.

DATA SOURCES

Guidelines, protocols, and published articles.

CONCLUSIONS

There is a lack of scientific basis for food restrictions, a wide variation in policies related to low-microbial diets, and inconsistent compliance with restricted diets. Furthermore, lack of consistent practice has not been related to incidence of infection.

IMPLICATIONS FOR NURSING PRACTICE

Evaluation of an evidence-based project such as this can lead to change in practice and institutional policy.

Nosocomial Pneumonia is Not Prevented by Protective Contact Isolation in the Surgical Intensive Care Unit

Nosocomial pneumonia (NP) is the leading cause of death from hospital-acquired infection in intubated surgical intensive care unit (SICU) patients. To determine whether protective contact isolation would lower the incidence of NP in intubated patients we performed a prospective, randomized, and controlled study in two SICUs in a tertiary medical center. Over a period of 15 months two identical ten-bed SICUs alternated for 3-month periods between protective contact isolation (isolation group) and standard “universal precautions” (control group). In the isolation group all personnel and visitors donned disposable gowns and nonsterile gloves before entering an intubated patient's room; handwashing was required before entry and on leaving the room. In the control group caregivers utilized only “standard precautions” including handwashing and nonsterile gloves for intubated patients. Respiratory cultures were obtained 48 hours after SICU admission and every 48 hours thereafter until extubation, transfer to floor care, or death. Airway colonization (AC) occurred in 72.7 per cent of isolated patients and 69.0 per cent of control patients ( P = 0.61). The incidence of NP was significantly higher in the isolation group (36.4%) compared with the control group (19.5%) ( P = 0.02). There was no statistically significant difference between groups in days from SICU admission to AC, days to NP, and mortality. We conclude that protective contact isolation with gowns, gloves, and handwashing is not superior to gloves and handwashing alone in the prevention of AC and NP in SICU patients and may in fact be detrimental.

Infection control in the ICU

Nosocomial infections (NIs) now concern 5 to 15% of hospitalized patients and can lead to complications in 25 to 33% of those patients admitted to ICUs. The most common causes are pneumonia related to mechanical ventilation, intra-abdominal infections following trauma or surgery, and bacteremia derived from intravascular devices. This overview is targeted at ICU physicians to convince them that the principles of infection control in the ICU are based on simple concepts and that the application of preventive strategies should not be viewed as an administrative or constraining control of their activity but, rather, as basic measures that are easy to implement at the bedside. A detailed knowledge of the epidemiology, based on adequate surveillance methodologies, is necessary to understand the pathophysiology and the rationale of preventive strategies that have been demonstrated to be effective. The principles of general preventive measures such as the implementation of standard and isolation precautions, and the control of antibiotic use are reviewed. Specific practical measures, targeted at the practical prevention and control of ventilator-associated pneumonia, sinusitis, and bloodstream, urinary tract, and surgical site infections are detailed. Recent data strongly confirm that these strategies may only be effective over prolonged periods if they can be integrated into the behavior of all staff members who are involved in patient care. Accordingly, infection control measures are to be viewed as a priority and have to be integrated fully into the continuous process of improvement of the quality of care.

[Febrile neutropenia: practical aspects]

BACKGROUND

Infections are a major cause of mortality in neutropenic patients. They require long hospital stays and highly expensive therapeutic measures. In this review we discuss the practical and pharmaco-economic aspects of the management of febrile neutropenia.

PREVENTION AND THERAPY

Prevention of fever of unknown origin (FUO) demands hygienic and antimicrobiotic measures. First-line antibiotic therapy consists of an aminoglycoside combined with an ureidopenicillin or a 3rd-generation cephalosporin. Double beta-lactam antibiotic combinations are equally effective and less toxic, but more expensive. Monotherapy with carbapenems, ceftazidime, or cefepime appear to offer comparable efficacy. Lung infiltrates require immediate treatment with amphotericin B. If the initial therapeutic regime fails, a carbapenem plus a glycopeptide antibiotic and a parenteral antimycotic drug should be applied after 3 to 4 days. The prophylactic or interventional administration of hematopoietic growth factors is only indicated in special high-risk situations.

CONCLUSIONS

Using the described therapeutic procedure, the response rate exceeds 90%. Consistent, step-wise escalating administration of antibiotics is essential. More evaluation is needed to determine whether selected patients with febrile neutropenia can be treated on an outpatient basis.

Prevalence of use of infection control rituals and outdated practices: Education Committee survey results

BACKGROUND

To better serve APIC membership, the National APIC Education Committee periodically conducts educational needs assessments.

METHODS

In the July/August 1995 issue of APIC News, the committee published a survey on the prevalence of infection control "sacred cows," practices considered to be outdated and ritualistic. A response was requested only from ICPs working in inpatient facilities. We asked ICPs which of the listed practices were part of their facility's infection control program, and whether they were interested in changing those that were. We also asked about the certification status of the program's manager (CIC vs non-CIC).

RESULTS

A total of 506 survey forms were completed and returned, with 74% from acute care only, 7% from long-term care only, 10% from acute and long-term care, and 8% from other types of facilities. Mean number of beds was 226. Infection control programs had existed for a mean of 15 years. Most respondents' facilities still engaged in infection control rituals. Fifty-eight percent of respondents were still doing total surveillance, 43% were still using reverse or protective isolation, and 11% were still subjecting employees with positive reactions to purified protein derivative of tuberculin testing to annual chest radiographs, to name a few. Certified respondents were significantly less likely to include outdated practices than were non-CICs for 11 of 15 practices (73%). CICs were no more likely than non-CICs, however, to be interested in changing any current rituals.

CONCLUSION

The Education Committee will use this information to plan future educational offerings.

Selective discharge of patients with acute myeloid leukemia during chemotherapy-induced neutropenia

PURPOSE

It is common practice for patients with acute myeloid leukemia (AML) to be observed in hospital during the entire nadir after intensive chemotherapy. In an attempt to lessen the likelihood of developing infections with hospital acquired pathogens, we usually discharge patients upon completion of chemotherapy and follow them as outpatients. They are readmitted if fever develops. We evaluated the feasibility and safety of this practice.

PATIENTS AND METHODS

We studied 29 patients with AML (median age 40 years, range 16-63) who were treated with intensive remission-induction and consolidation chemotherapy. Afebrile patients not receiving antibiotics were discharged immediately following chemotherapy and were followed every 3-4 days at the day care unit. Patients were instructed to return immediately if fever rose to 38.2 degrees C or a fever of 38 degrees C persisted for 2 hr. The 29 patients received a total of 86 courses. Following 50 courses, patients were discharged. These 50 ambulatory nadir periods (ANPs) were monitored.

RESULTS

Median WBC and platelet counts on discharge were 2,900 per cubic millimeter (range 300-8,300) and 137,000 per cubic millimeter (range 17,000-618,000), respectively. Mean traveling time from the hospital by car was 1.6 hr (range 15 min-3 hr). In three of the 50 ANPs (6%), patients were not readmitted during their entire nadir. During 47 of the ANPs, patients returned to the hospital (because of fever in 44 cases), a mean of 7.2 days (range 1.0-12.7 days) after discharge. In 45 ANPs, patients were readmitted in good general condition. Four patients had life-threatening complications. Two patients were admitted in septic shock due to delay in seeking admission, but rapidly recovered. Two other patients died, one of cardiogenic shock within 24 hr of readmission and one 24 days later. Only one of the 11 gram negative bacteria cultured was resistant to mezlocillin and gentamicin. After 45 ANPs, patients were discharged a mean of 12.2 days (range 5-42 days) following readmission. We estimate that approximately 383 hospital days were saved by this policy, a mean of 7.6 days per patient, representing 16% of total inpatient hospital days.

CONCLUSIONS

For AML patients who are reliable and without complicating medical conditions, selected discharge following chemotherapy is a low-risk practice and may reduce the incidence of infection with resistant hospital-acquired pathogens.

Antibiotic-resistant gram-negative bacteria in the critical care setting

Gram-negative bacilli that are resistant to commonly used antibiotics are a growing problem in seriously ill, hospitalized patients. Numerous outbreaks involving these organisms have been reported in intensive care nurseries and among critically ill adults. In endemic situations, the major reservoir for these pathogens is the patient; occasionally, transmission from patient to patient occurs through the hands of caregivers. Although the degree of antibiotic use probably plays some role in the emergence of antibiotic-resistant gram-negative bacilli, this relationship has not been uniformly demonstrated, and other factors intrinsic to the organisms themselves and to the critically ill patient may play an important role.

Intraperitoneal administration of leukotriene B4 (LTB4) and omega-guanidino caproic acid methane sulfonate (GCA) increased the survival of mice challenged with methicillin-resistant Staphylococcus aureus (MRSA)

Infections caused by methicillin-resistant Staphylococcus aureus (MRSA) very often complicate management of immunocompromised patients. We studied the effect of leukotriene B4 (LTB4) and epsilon-guanidino caproic acid methane sulfonate (GCA), on MRSA infection. Mice fed a 20% casein diet were intraperitoneally administered LTB4, GCA, or saline (control) daily for 30 days. On the 10th day of this treatment, mice were challenged with MRSA. The survival rate in the control group (20%) was significantly lower than the rates in the GCA (60%) and LTB4 (50%) groups, respectively (p < 0.05). There was a significant reduction of MRSA in the spleen and kidney of the survived mice in GCA group as against mice in the LTB4 and saline groups, indicating a better recovery in GCA group than the other groups. The results suggest that intraperitoneal administration of GCA and LTB4 may play a role in host defense mechanism during MRSA infections.

Use of the quinolones for the prophylaxis and therapy of infections in immunocompromised hosts

The prevention and treatment of infections are major issues of supportive care in patients with haematological malignancies. Because of their broad antimicrobial activity, the use of fluoroquinolones for prophylaxis in neutropenic patients has been extensively studied. In comparison with placebo, norfloxacin reduces the incidence of Gram-negative infections, whereas Gram-positive bacterial and fungal infections remain unaffected. Ofloxacin and enoxacin also bacterial and fungal infections remain unaffected. Ofloxacin and enoxacin also produce a reduction in fever and documented infections. In randomized studies comparing ciprofloxacin with cotrimoxazole (trimethoprim/sulfamethoxazole) plus colistin (each in combination with nonabsorbable antifungal agents), conflicting results were obtained. The incidence of documented Gram-negative bacterial infections was markedly reduced by ciprofloxacin prophylaxis; however, the number of Gram-positive infections may increase dramatically. Combining ciprofloxacin with a macrolide antibiotic in an attempt to prevent streptococcal infections can result in breakthrough bacteraemias due to resistant Gram-positive pathogens. Empirical antimicrobial therapy after quinolone prophylaxis should also be directed against microorganisms susceptible to quinolones, since sustained eradication by oral administration cannot be assumed with certainty. Clinical trials comparing intravenous quinolones in combination with aminoglycosides with widely used standard regimens for the treatment of infections in cancer patients indicate equivalent efficacy; however, in studies of ciprofloxacin alone, response rates were significantly lower compared with standard combinations. Therefore, quinolone monotherapy as empirical treatment in febrile neutropenic patients cannot be recommended.

An outbreak of multiply resistant Pseudomonas aeruginosa in a neonatal unit: plasmid pattern analysis

An outbreak of infection due to multiply resistant Pseudomonas aeruginosa occurred from March to April 1986 in a neonatal unit. Affected neonates were receiving ventilation support and the mortality rate was high. Plasmid analysis and antibiograms indicated that the outbreak was due to a single strain. A survey of bacteria isolated from respirators, potable water and hands of personnel working in the unit failed to recover the outbreak strain. Lack of sterilization of respirators and overcrowding were considered to be the causes of the outbreak and reinforcement of the importance of aseptic techniques helped in its termination.

The value of protective isolation procedures in cardiac allograft recipients

The impact of protective isolation on the incidence of infection in 60 cardiac transplant recipients (mean age, 49.2 years) was studied in a prospective randomized trial. Thirty patients were randomized to protective isolation, which consisted of private room, hat, mask, sterile gown, and handwashing. Thirty patients were randomized to no isolation, which meant they recovered in a crowded, open intensive care unit and were adjacent to recipients of liver transplants or patients who were on the trauma, neurosurgical, and general surgical services, many of whom had an infection of the incision or a pulmonary infection. There was no difference between groups in the proportion of patients in whom infection developed (chi 2[1] = 0.27; p = 0.6), the number of infection-related deaths (2 in each group), the types of infection (bacterial, viral, fungal, or protozoal), or the overall outcome. Because protective isolation offered no benefit over standard care in protecting these patients from infections or the associated complications, we have discontinued its routine use after cardiac transplantation.

Environmental control in management of immunodeficient patients: experience with "David"

Environmental control in managing patients with immunodeficiency ranges from the exceedingly complex to the relatively simple. At one end of the spectrum is the total isolation technology applied to David, the "Bubble Boy" who lived his entire life behind sterile plastic barriers. At the other end of the spectrum is the simpler technology applied to patients receiving bone marrow transplants who are maintained in ordinary private hospital rooms and attended by personnel who merely observe handwashing precautions. Most properly performed and controlled studies of the use of special isolation procedures to reduce infections derive from patients receiving bone marrow transplants for conditions of aplastic anemia and leukemia or patients receiving chemotherapy for malignancy. The design of isolation procedures for immunodeficient patients borrows from these studies because of the relatively small number of immunodeficient patients. These studies have shown that laminar airflow rooms produce a significantly lower incidence of infections but may not change the mortality of all patients. Also, protective isolation has clearly reduced the incidence and severity of graft-versus-host disease in transplanted patients with aplastic anemia. Recently there has been a trend away from strict isolation procedures because careful studies have indicated that host rather than acquired pathogens are responsible for at least 85% of infections in these special patients. Also, the human stress of prolonged isolation is becoming increasingly recognized. The complex and expensive isolation techniques that were used in David's case are no longer being utilized in immunodeficient subjects, partly because new transplantation technology has made it possible to cross histocompatibility barriers, obviating the need for permanent isolation.

Cost-effective application of the Centers for Disease Control Guideline for Isolation Precautions in Hospitals

Isolation, as practiced in many hospitals today, is associated with ineffective, costly rituals. The isolation process can be monitored and critically examined to determine illogical practices. When these practices and their associated costs are known, specific strategies can be implemented to reduce the costs associated with isolation.

Nosocomial infections

Recent developments in the understanding of nosocomial infection in general and nosocomial respiratory infections in particular are elucidated. Although the discussion focuses on aerobic bacteria, data are presented to indicate the growing realization that unusual and newly discovered microorganisms play a significant role in hospital-acquired infections. Strategies for the control or prevention of nosocomial infections are highlighted.

The use of sterile and low microbial diets in ultraisolation environments

The evidence for the use of sterile and low microbial diets in ultraisolation environments is reviewed. Studies have suggested that sterile food is not required for gut sterilization when oral nonabsorbable antibiotics are used, but if a low microbial food contains an antibiotic-resistant organism, colonization can occur. There may be a beneficial effect on the incidence of infection by serving pathogen-free foods, either sterile or low microbial, to the immunosuppressed patient regardless of type of environment, yet the comparative effectiveness of sterile and low microbial diets in preventing introduction of new pathogens accessing the host via the mouth, oropharynx, and esophagus has not been systematically evaluated.

Double-blind randomized study of prophylactic trimethoprim/sulfamethoxazole in granulocytopenic patients with hematologic malignancies

In a double blind study, oral prophylactic trimethoprim/sulfamethoxazole was evaluated for its utility in preventing serious infections in patients with hematologic malignancy. Of 58 evaluated granulocytopenic episodes in 47 patients, acute leukemia was the underlying malignancy in 46 episodes. Trimethoprim/sulfamethoxazole prophylaxis resulted in fewer microbiologically documented infections (seven versus 15; p = 0.029). This was primarily the result of a reduction in episodes of bacteremia in the trimethoprim/sulfamethoxazole-treated group as compared with the placebo-treated group (three versus nine episodes; p = 0.05). The combined frequency of disseminated candidiasis, candidemia, and esophagitis of presumed fungal etiology was greater in the trimethoprim/sulfamethoxazole-treated group (six) than in the placebo-treated group (two) but not significantly so (p = 0.13). Similarly, there were no significant differences between groups in the overall incidence of infectious complications, number of febrile days, use of parenteral antibiotics, or number of days following randomization to first infectious episode. Throat and rectal surveillance cultures more frequently revealed trimethoprim/sulfamethoxazole-resistant gram-negative bacilli and yeasts in the trimethoprim/sulfamethoxazole-treated group. More frequent emergence of yeast isolates from previously culture-negative patients was documented (p = 0.033). Thus, in this study, trimethoprim/sulfamethoxazole prophylaxis during granulocytopenia reduced the incidence of microbiologically documented infections. However, the emergence of resistant bacteria and of fungi may limit the potential usefulness of this approach.

Oral antibiotic prophylaxis in patients with cancer: a double-blind randomized placebo-controlled trial

In an attempt to reduce the incidence of fever and infection, we randomized patients with cancer to receive trimethoprim/sulfamethoxazole plus erythromycin (TMP/SMX + E) versus placebos after each cycle of chemotherapy (no crossover) and to continue until granulocytopenia (polymorphonuclear leukocytes less than 500/mm3) resolved or the patient became febrile. We evaluated 541 episodes (150 patients); 249 episodes (77 patients) with TMP/SMX + E and 292 episodes (73 patients) with placebos. The patients' median age was 17 years. Thirty percent of the patients had leukemia, 23% had lymphoma, and 47% had solid tumors. Compliance with prescribed medication was prospectively rated as excellent in 60.6%, good in 11.7%, poor in 11.1%, and unknown in 16.6%; compliance was better for the placebo group (P = 0.001). The overall incidence of fever or infection was 22.1% for the TMP/SMX + E group versus 26.9% for the placebo group. When only episodes with excellent compliance in which granulocytopenia was documented were compared, the incidence of fever or infection was 18.1% for the TMP/SMX + E group vs 32.2% for the placebo group (P = 0.009), with bacterial infection occurring in 3.8% of the TMP/SMX + E group vs 11.9% of the placebo group (P = 0.019), and unexplained fever in 10.5% of the TMP/SMX + E group vs 19.6% of the placebo group (P = 0.037). Patients with good or poor compliance showed no significant benefit from the TMP/SMX + E, and patients with excellent compliance did best, regardless of whether they were receiving antibiotics or placebos, suggesting that patient compliance is an important independent variable. The incidence of fever or infection was significantly lower for patients with leukemia with excellent compliance who received antibiotics (P = 0.037) than for patients with lymphomas or solid tumors. Oral antibiotic prophylaxis reduced the incidence of fever and infection in some granulocytopenic patients, but the benefit was limited and restricted to patients whose compliance was complete.