Low bacterial diet to prevent infection in neutropenic patients
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To determine the efficacy of LBD in patients with compromised immune systems during the episode of neutropenia to prevent Gram‐negative or Gram‐positive infections, fungaemia and related mortality.
Background
Severe neutropenia, defined as Absolute Neutrophil Count (ANC) of < 0.5 x 109 /L, is a potentially critical side effect of chemotherapy and is the major risk factor for infection and sepsis due to substantial suppression of the immune system. The neutrophil, constituting 55% to 70% of the circulating white blood cells (WBC), has the ability to seek out, ingest and destroy the majority of foreign invaders (Candell 1991). Patients with both solid tumours and haematological malignancies treated with high dose chemotherapy have a significantly increased risk of developing life‐threatening infections. When the neutrophil count falls from 2 to 7 x 109 /L to less than 0.5 x 109 /L (neutropenia) there is an increased susceptibility to infection. Neutropenic Human Immunodeficiency Virus (HIV) infected patients and patients with Acquired Immune Deficiency Syndrome (AIDS) are also significantly at risk for developing severe infections (Hermans 1999). The frequency and severity of infections is inversely proportional to the neutrophil count and directly proportional to the duration of neutropenia (Hughes 2002).
Approximately 70% to 75% of deaths from acute leukaemia and 50% of deaths from solid tumours are related to infection secondary to neutropenia (Barber 2001). For neutropenic patients with AIDS the adjusted relative risk (RR) for the occurrence of bacterial infection is 2.33 (Moore 1996). At least 50% of neutropenic patients who become febrile have an established or occult infection and at least one‐fifth of patients with a neutrophil count of less than 0.1 x 109/L have bacteraemia (Hughes 2002). Approximately 80% of the organisms causing infections in the neutropenic patient arise from endogenous microbial flora colonizing the skin and respiratory, genitourinary and gastrointestinal tracts. Gram‐negative organisms such as Enterobacter or Pseudomonas are the most frequent and serious causes of infection. Gram‐positive organisms like Staphylococcus, Streptococcus, Corynebacteria and Clostidia have an increased incidence because of the use of central venous catheters (Barber 2001). Invasive fungal infections are also important causes of morbidity and mortality. Predisposing factors for fungal infections include broad‐spectrum antibiotics, corticosteroids, parenteral nutrition indwelling intravenous catheters and graft‐versus host disease after an allogeneic Stem Cell Transplantation. Most commonly isolated fungal pathogens are Aspergillus and Candida (Barber 2001)
Significant advances in supportive care for neutropenic patients have been made since the mid‐1990s. Nowadays the supportive care management for neutropenia is led by risk‐assessments (Klastersky 2000; Talcott 1992) and guidelines (Dykewicz 2001; Hughes 2002). The Centre of Disease Control developed evidence‐based guidelines for the management of neutropenia and preventing opportunistic infections (Dykewicz 2001). These recommendations to prevent healthcare‐associated infections concern the use of antimicrobial prophylaxis, colony stimulating factors, protective environment, oral care, central venous catheter care, hand washing, personal hygiene practices, dietary restrictions and out patient treatment (Dykewicz 2001). However, despite these achievements infection continues to be the major cause of morbidity and mortality in the neutropenic patient.
With regards to dietary restriction it has been hypothesized that a neutropenic diet reduces pathogens in the gastrointestinal tract by excluding specific foods that may act as vectors of bacteria. The first intervention began in the 1960s with the intention of providing a completely sterile diet (Reimer 1966). Foods were sterilised by autoclaving, prolonged baking, gamma irradiation or canning with the aim of producing food that has no bacterial or fungal growth (Aker 1983). Since sterile diets were considered unpalatable by patients, the US National Institutes of Health, Department of Dietary and Environmental Sanitation, designed the 'cooked food' diet. Although not sterile, this diet was screened to eliminate foods with high bacterial counts (Preisler 1970). In a randomized trial, the National Cancer Institute demonstrated that within a decontaminated environment, the sterile diet gave little advantage over the cooked food diet with reference to bacterial stool cultures (Preisler 1970). Although the cooked food diet was more acceptable to patients than the sterile diet, patients who adhered to this diet for longer than 4 to 6 weeks often became frustrated with the food selection (Moody 2002). Occasionally this diet affected their acceptance of other medical therapies as well which led clinicians to investigate liberalization of the diet (Pizzo 1982). Pizzo et al cultured 236 commercially available foods and found that 66% of them grew less than 500 colony‐forming units/gram of food in microbiologic studies. Therefore, they proposed that these foods were acceptable for neutropenic patients. This liberalized diet became known as the Low Bacterial Diet (LBD) (Pizzo 1982).
The role of diet in the development of infection in patients with neutropenia is controversial (French 2001). Dietary restrictions vary in the literature and amongst institutions. Recommendations range from no dietary restrictions to extensive restrictions. Two surveys (French 2001; Smith 2000) indicated that several differences existed with the LBDs offered by hospitals in the USA. Furthermore there was much variation regarding the initiation and discontinuation point of the LBD. Few clinical studies have been done to assess the efficacy of the LBD in reducing infection rates in neutropenic patients and currently there is no substantial evidence to prove the benefit of the LBD (Larson 2004). As it may pose an unnecessary burden for the patient who already has difficulty with maintaining an adequate oral diet due to factors related to high dose treatment, it is clear that it would be beneficial to expand the body of knowledge regarding the effectiveness of LBD. A systematic review in this area up to date has not been performed; therefore this systematic review could provide relevant information on the efficacy of LBD in reducing morbidity and mortality.
Our aim in this systematic review is to evaluate the efficacy of LBD in decreasing bacteraemia and infection related mortality in neutropenic patients. Furthermore, it is also a possibility to shed light on this complex issue and to propose area of further investigation in case of unsolved questions.
Objectives
To determine the efficacy of LBD in patients with compromised immune systems during the episode of neutropenia to prevent Gram‐negative or Gram‐positive infections, fungaemia and related mortality.
Methods
Criteria for considering studies for this review
Types of studies
Randomized controlled trials (RCTs) comparing use of LBD versus no use of LBD. Non randomised studies (quasi‐RCTs, non‐randomised controlled trials, cohort studies, case control studies) will only be included in the absence of RCTs due to their susceptibility to bias.
Types of participants
Patients (both adults and children aged one year and above) with compromised immune systems (acquired or congenital or patients with cancer) who are likely to receive chemotherapy causing episodes of neutropenia. Both adults and children will be included, as the adult patient is comparable to the paediatric patient as far as interpreting infections in the gastro‐intestinal tract.
Types of interventions
Intervention and control arm
-
Intervention: use of LBD
-
Control arm: no use of LBD
LBD is defined as: A diet intended to reduce the ingestion of bacterial and fungal contaminants by the exclusion of foods such as uncooked fruits and vegetables, cold cuts, undercooked eggs and meat, unsterilized water, unpasteurized milk products and soft cheeses. These food restrictions are highly variable.
Types of outcome measures
Primary outcomes
The primary outcome is clinically documented infection during neutropenic episodes (Gram‐negative or Gram‐positive bacteraemia, fungaemia, proven fungal infection) and mortality (infection related mortality)
(For definitions see Table 1)
| Infection is defined as: |
Secondary outcomes
-
Time to first febrile episode of fever
-
Need for empirical antibiotic therapy (Feld 2002)
-
Quality of life (QOL) to accept LBD
Search methods for identification of studies
Electronic searches
We will search the Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 3, 2006). MEDLINE, EMBASE and CINAHL will be searched for the years 1966 to 2006. More specifically the Cochrane Database of Systematic Reviews, the Database of Abstracts of Reviews of Effects (DARE) also will be searched.
The following search terms will be used: (adults, children)
agranulocytosis OR bacterial translocation OR immunocompromised host OR neutropeni* OR leukopeni* OR leucopeni* OR granulocytopeni* OR immunocompromised OR
immunocompromised) AND ((low OR minimal OR "germ poor" OR neutropenic OR cooked
OR "reduced bacteria*") AND (diet OR feeding OR diet* OR food* OR nutrition) OR "dietary restriction*"
NOT (animals [mn] NOT human [mh]
Information about trials not registered in MEDLINE, CINAHL, EMBASE, either published or unpublished, will be located by searching the reference lists of relevant articles and review articles.
Searching other resources
We will hand search the following conference proceedings: ASH 2000 to 2005 (American Society of Hematology), EBMT 2000‐2005 (European Bone and Marrow Transplantation), SIOP 2000 to 2005 (International Society for Paediatric Oncology), MASCC 2000 to 2005 (Multinational Association of Supportive Care in Cancer), ASCO 2000 to 2005 (American Society of Clinical Oncology), ICAAC 2000 to 2005 (Interscience Conference of Antimicrobial agents and Chemotherapy), ESPEN 2000 to 2005 (European Society for Clinical Nutrition and Metabolism), ASPEN 2000 to 2005(American Society for Parenteral and Enteral Nutrition), EHA 2000‐2005 (European Hematology Association). Researchers working in this area will be contacted to enable identification of ongoing trials. Language restriction will not be imposed. The searches will be updated every two years.
Data collection and analysis
Selection of studies
All titles and abstracts retrieved by electronic and hand searching will be downloaded to a reference management database (e.g. Reference Manager or Endnote) and duplicates will be removed. This will facilitate the production of a Quorom flow diagram (using the Additional figures facility of RevMan) showing the total number of potential studies identified, the number excluded at each stage, and the final number meeting the inclusion criteria (Moher 1999). Identification of studies meeting the eligibility criteria will be performed independently by three authors. Three authors (AM, NL and MD) will independently screen the titles and abstracts of references identified by the search for potential relevance. Disagreement or doubt will be resolved by discussion. Any study which seems to meet the inclusion criteria on grounds of the title, abstract or both will be obtained and studied in detail. Information regarding the reasons for exclusion of any study considered for inclusion will be clearly stated. When no RCTs which meet the inclusion criteria can be found the level of evidence will be decreased to non‐RCTs, cohort and case control studies.
Data extraction and management
Data extraction will be performed independently by three authors (AM, NL, MD) using standardised forms. Data of the characteristics of the participants (age, sex, disease, treatment, antimicrobial prophylaxis, colony stimulating factors, protective environment, oral care, central venous catheter care, hand washing, hygiene practices), interventions (use of LBD or not), will be extracted. For binary outcomes (e.g. infection, fungaemia, time to first febrile episode of fever, need for empirical antibiotic therapy, QOL to accept LBD and infection and fungal related mortality). We will abstract the number of patients randomised to each treatment arm and the number who experienced the outcome of interest, in order to estimate a relative risk (RR). Where possible, the outcome data abstracted will be those required for an intention‐to‐treat analysis. In case of disagreement, the abstract and articles will be re‐examined and discussed until consensus is achieved. If this is not impossible, a fourth author will be consulted.
Assessment of risk of bias in included studies
Studies fulfilling the review inclusion criteria will be assessed for methodological quality by three review authors (AM, NL and MD) working independently. The quality of RCTs will be assessed using the criteria described in the Cochrane Handbook (Higgins 2005) (Table 2).
| Selection bias |
For assessing the methodological quality of non‐randomised studies the Newcastle‐Ottawa assessment tool will be used (Wells 2006) (Table 3). This tool was developed for quality assessment of non‐randomised studies to be used in a systematic reviews and contains eight items categorised into the three groups: selection, comparability and outcome. Possible confounding variables will be identified for all studies by a 'star system'. This means the more confounding variables identified, the more stars they will get.
| NEWCASTLE ‐ OTTAWA QUALITY ASSESSMENT SCALE |
Discrepancies between review authors will be solved by discussion, if necessary a fourth review author will be consulted.
Data synthesis
Meta‐analysis will be performed using the Cochrane Review Manager 4.2. For binary outcomes (e.g. death, infection, colonisation), the RR will be calculated for each study and these statistics will be pooled in meta‐analyses. Randomized and observational (non‐randomized) studies will be analyzed separately as non‐randomized comparisons can overestimate treatment effects. Among observational studies, cohort and case‐controlled studies will be analyzed separately. Data will be pooled. If patients are enrolled more than once in a period of neutropenia separate analysis will be performed (Feld 2002).
All data will be analysed with a fixed effect model. The significance of heterogeneity between trial results will be tested by using a standard chi‐squared test. The magnitude of heterogeneity between trial results will be quantified using inconsistency the I2 statistic (Higgins 2003), which describes the percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error. A value greater than 50% will be considered substantial heterogeneity. Possible causes of heterogeneity will be explored by conducting subgroup and sensitivity analyses.
| Infection is defined as: |
| Selection bias |
| NEWCASTLE ‐ OTTAWA QUALITY ASSESSMENT SCALE |
