Urinary catheter policies for long-term bladder drainage

  • Review
  • Intervention

Authors


Abstract

Background

People requiring long-term bladder draining commonly experience catheter-associated urinary tract infection and other problems.

Objectives

To determine if certain catheter policies are better than others in terms of effectiveness, complications, quality of life and cost-effectiveness in long-term catheterised adults and children.

Search methods

We searched the Cochrane Incontinence Group Specialised Trials Register (searched 28 September 2011). Additionally, we examined all reference lists of identified trials.

Selection criteria

All randomised and quasi-randomised trials comparing catheter policies (route of insertion and use of antibiotics) for long-term (more than 14 days) catheterisation in adults and children.

Data collection and analysis

Data were extracted by two reviewers independently and compared. Disagreements were resolved by discussion. Data were processed as described in the Cochrane Handbook. If the data in trials had not been fully reported, clarification was sought from the authors. When necessary, the incidence-density rates (IDR) and/or the incidence-density differences (IDD) within a certain time period were calculated.

Main results

Eight trials met the inclusion criteria involving 504 patients in four cross-over and four parallel-group randomised controlled trials. Only two of the pre-stated six comparisons were addressed in these trials.

Four trials compared antibiotic prophylaxis with antibiotics when clinically indicated. For patients using intermittent catheterisation, there were inconsistent findings about the effect of antibiotic prophylaxis on symptomatic urinary tract infection (UTI). Only one study found a significant difference in the frequency of UTI favouring prophylaxis. For patients using indwelling urethral catheterisation, one small trial reported fewer episodes of symptomatic UTI in the prophylaxis group.

Four trials compared antibiotic prophylaxis with giving antibiotics when microbiologically indicated. For patients using intermittent catheterisation, there was limited evidence that receiving antibiotics reduced the rate of bacteriuria (asymptomatic and symptomatic). There was weak evidence that prophylactic antibiotics were better in terms of fewer symptomatic bacteriuria.

Authors' conclusions

No eligible trials were identified that compared alternative routes of catheter insertion. The data from eight trials comparing different antibiotic policies were sparse, particularly when intermittent catheterisation was considered separately from indwelling catheterisation. Possible benefits of antibiotic prophylaxis must be balanced against possible adverse effects, such as development of antibiotic resistant bacteria. These cannot be reliably estimated from currently available trials.

Résumé scientifique

Règles de manipulation d’un cathéter urinaire pour le drainage de la vessie à long terme

Contexte

Les personnes nécessitant un drainage à long terme de la vessie présentent couramment une infection urinaire liée au cathéter et autres problèmes.

Objectifs

Déterminer si certaines règles de manipulation des cathéters sont meilleures que d'autres en termes d'efficacité, de complications, de qualité de vie et de rentabilité chez les adultes et les enfants portant un cathéter à long terme.

Stratégie de recherche documentaire

Nous avons effectué des recherches dans le registre d’essais spécialisés du groupe Cochrane sur l’incontinence (recherches effectuées le 28 septembre 2011). Par ailleurs, nous avons examiné toutes les listes bibliographiques des essais identifiés.

Critères de sélection

Tous les essais randomisés et quasi-randomisés comparant les règles de manipulation des cathéters (voie d'insertion et utilisation d'antibiotiques) pour le cathétérisme à long terme (plus de 14 jours) chez les adultes et les enfants.

Recueil et analyse des données

Les données ont été extraites par deux évaluateurs de manière indépendante et comparées. Des discussions ont permis de résoudre des désaccords. Les données étaient traitées comme indiqué dans le manuel Cochrane. Si les données des essais n'avaient pas été entièrement présentées, une clarification était demandée aux auteurs. Si nécessaire, les densités d’incidence (DI) et/ou les différences de densité d'incidence (DDI) sur une certaine période étaient calculés.

Résultats principaux

Huit essais, portant sur 504 patients dans quatre essais contrôlés randomisés croisés et en groupes parallèles, répondaient aux critères d'inclusion. Seuls deux des six comparaisons prémentionnées étaient étudiées dans ces essais.

Quatre essais comparaient la prophylaxie antibiotique aux antibiotiques en cas d'indication clinique. Pour les patients utilisant un cathétérisme intermittent, les résultats relatifs à l'effet de la prophylaxie antibiotique sur l'infection urinaire (IU) symptomatique étaient contradictoires. Seule une étude a trouvé une différence significative de fréquence de l'IU favorisant la prophylaxie. Pour les patients utilisant un cathétérisme urétral permanent, un seul essai de petite taille a rapporté un nombre inférieur d'épisodes d'IU symptomatique dans le groupe de prophylaxie.

Quatre essais comparaient la prophylaxie antibiotique à l'administration d'antibiotiques en cas d'indication microbiologique. Pour les patients utilisant un cathétérisme intermittent, il existait des preuves limitées que le traitement antibiotique réduisait le taux de bactériurie (asymptomatique et symptomatique). Il existait des preuves peu probantes que les antibiotiques prophylactiques étaient plus efficaces en termes de réduction du taux de bactériurie symptomatique.

Conclusions des auteurs

Aucun essai éligible comparant des modes alternatifs d'insertion des cathéters n'a été identifié. Les données provenant des huit essais comparant les différentes politiques d'administration d'antibiotiques, en particulier en cas de cathétérisme intermittent, ont été examinées séparément du cathétérisme permanent. Les éventuels bénéfices de la prophylaxie antibiotique doivent être comparés aux éventuels effets indésirables, tels que le développement de bactéries résistantes aux antibiotiques. Il est impossible d'évaluer ceux-ci avec fiabilité d'après les essais actuellement disponibles.

Plain language summary

Urinary catheter policies for long-term bladder drainage

Some people use catheters to help them manage their bladder problems (such as leaking urine or not being able to pass urine). Catheters may be permanent urethral catheters (in the tube draining the bladder), suprapubic catheters (via the abdomen) or intermittent catheters (when a catheter is inserted via the urethra several times a day). No trials were found comparing these different methods with each other. Sometimes people using the catheters develop urinary tract infections. There was some weak evidence that using antibiotics all the time reduced the chance of having a urinary tract infection while using intermittent catheters, but there was not enough information about side effects.

Résumé simplifié

Règles de manipulation d’un cathéter urinaire pour le drainage de la vessie à long terme

Certaines personnes utilisent un cathéter qui les aide à gérer leurs problèmes de vessie (tels que les fuites d'urine ou l'incapacité à uriner). Les cathéters peuvent être des cathéters urétraux permanents (dans le tube drainant la vessie), des cathéters suspubiens (via l'abdomen) ou des cathéters intermittents (lorsqu'un cathéter est inséré par l'urètre plusieurs fois par jour). Aucun essai comparant ces différentes méthodes les unes aux autres n'a été trouvé. Il arrive que les personnes qui utilisent des cathéters développent des infections urinaires. Des données peu probantes semblaient indiquer que l'utilisation permanente d'antibiotiques réduisait le risque de présenter une infection urinaire en cas d'utilisation intermittente de cathéters, mais les données sur les effets secondaires étaient insuffisantes.

Notes de traduction

Traduit par: French Cochrane Centre 13th September, 2012
Traduction financée par: Ministère du Travail, de l'Emploi et de la Santé Français

Background

Description of the condition

Long-term catheterisation of the bladder is most commonly used in the care of frail, elderly people in nursing homes or as part of their home care. In Canada, about 5% to 10% of residents of long-term care facilities have chronic indwelling catheters for weeks, months or years (Nicolle 2001). However, younger patients with neurological disease or spinal injury may also use catheters in the long term. The bladder performs two main functions, storage and emptying. The indications for catheterisation are failure to store and/or failure to empty, particularly urinary incontinence and urinary retention. In view of the ageing population in many countries, the number of people exposed to the risks of long-term indwelling catheterisation may increase, although the proportion of people whose urinary problems can only be managed by catheterisation may decrease because of improvements in other treatments for urinary incontinence and conditions which cause it.

Catheter-associated urinary tract infection represents the most common infection acquired in long-term facilities (Warren 1994). The incidence of bacteria in the urine (bacteriuria) has been estimated to be about 3% to 10% higher each day after catheter insertion (Warren 1982a). Therefore, after one month of catheterisation almost all patients would be expected to be bacteriuric. As in short-term catheterisation, asymptomatic bacteriuria can be complicated by the effects of symptomatic infection, e.g. fever, acute pyelonephritis and bacteraemia. Long-term catheterisation is also associated with catheter obstruction, urinary tract stones, local infections, chronic renal inflammation, renal failure and, over years, bladder cancer (Warren 1997).

Bacteria get into the catheterised bladder by direct inoculation at the time of catheter insertion or via the following routes during catheterisation: extraluminally by ascending from the urethral meatus along the catheter urethral interface, and intraluminally by reflux of the organisms into the catheter lumen (Tambyah 1999; Warren 2001). Biofilm formation and encrustation on the catheter surface, in combination with a high prevalence of recurrent bacterial entry, result in a dynamic polymicrobial bacteriuria (Liedl 2001).

Description of the intervention

The first step in reducing catheter-associated urinary tract infections and other complications is to avoid unnecessary catheterisation; the second is to remove the catheter as soon as possible. Alternative methods to indwelling urethral catheterisation should be considered, such as intermittent catheterisation and suprapubic catheterisation. From a theoretical point of view, it is plausible that both these methods are associated with a decreased risk of local infections and complications, but this remains to be proven.

Once a catheter is in place, the aim is to minimise the risk of infection. There are two accepted basic principles: keeping the catheter system closed; and removing the catheter when it is no longer needed. A further strategy concerns antibiotic prophylaxis: it is widely accepted that receiving systemic antibiotics reduces the risk of catheter-associated urinary tract infection in the short-term catheterised patient (Huth 1992; Platt 1986; Shapiro 1984). If systemic antibiotics are also effective in the long-term catheterised patient, we need to find out whether the protective effect of prophylactic antibiotics (used continuously) counterbalances the side effects, particularly the emergence of bacterial resistance to multiple antibiotics. Possible strategies to reduce these disadvantages are limiting antibiotic use: a) to patients with a positive urine culture (usually set at a predefined density of bacteria), b) to those with clinical symptoms (such as pain or fever), or c) to certain circumstances such as when changing the catheter.

How the intervention might work

With regard to intermittent catheterisation, i.e. inserting and removing a sterile or clean urethral catheter several times daily, micro-organisms are unlikely to gain entry into the bladder by ascending inside or along the outer surface of the catheter because the catheter is no longer constantly present. With respect to suprapubic catheterisation, bacterial colonisation of the urinary tract is less likely because of the lower density of (gram-negative) micro-organisms on the abdominal skin than in the periurethral area, although the catheter may still act as a route for infection. There are also disadvantages to these alternatives.

Why it is important to do this review

The first aim of this review is to address, in the context of long-term use, whether one of the catheter options (i.e. indwelling urethral catheterisation, intermittent urethral catheterisation or suprapubic catheterisation) is better than another in terms of suitability for the condition, complications, replacement, satisfaction (person with the problem, their carers and staff providing care) and cost-effectiveness. The second aim of this review is to assess the evidence from randomised controlled trials on the effectiveness of prophylactic (continuous) systemic (oral as opposed to local or topical) antibiotics, including which patient groups should receive antibiotics and in which circumstances, when a long period of catheterisation is anticipated.

Another strategy to reduce the incidence of problems in people who have long-term indwelling catheters is to wash out the bladder. However, this strategy was evaluated in another Cochrane review (Hagen 2010). With respect to indwelling catheterisation (i.e. urethral or suprapubic), a further possible strategy to reduce the risk of infection is to use catheters coated with an antibacterial substance. The effect of using different types of urethral catheters in the short-term (up to 14 days) is evaluated in a Cochrane review that is in the process of being updated (Schumm 2008b).

Objectives

To determine if certain catheter policies are better than others in terms of effectiveness, complications, quality of life and cost-effectiveness in long-term catheterised adults and children.
Specific comparisons to be addressed included:
1. indwelling urethral catheterisation compared with suprapubic catheterisation;
2. indwelling urethral catheterisation compared with intermittent catheterisation;
3. suprapubic catheterisation compared with intermittent catheterisation;
4. antibiotic prophylaxis compared with giving antibiotics when clinically indicated;
5. antibiotic prophylaxis compared with giving antibiotics when microbiologically indicated; and
6. giving antibiotics, if microbiologically indicated, compared with giving antibiotics if clinically indicated.

Methods

Criteria for considering studies for this review

Types of studies

All randomised and quasi-randomised trials comparing catheter policies (route of insertion and use of antibiotics) for long-term catheterisation of the bladder in adults and children.

Types of participants

All patients requiring long-term catheterisation for urinary incontinence or retention that cannot be managed by another method. This could include people suffering from stress, urge and mixed incontinence, dementia, prostatic hypertrophy unsuitable for other management, stroke, neurological problems, spinal cord injury and spina bifida. They may receive this care at home, in residential homes or in hospital. In this review, long term is defined as more than 14 days.

Types of interventions

The interventions considered were:
1. continuous indwelling urethral, intermittent urethral, and suprapubic use of catheters; and
2. antibiotic prophylaxis (continuous use), use of antibiotics if clinically indicated (e.g. pain, fever) and use of antibiotics if microbiologically indicated (growth of bacteria from a specimen of urine in the absence of clinical symptoms).

Antibiotics were divided into two categories:

  • broad spectrum (active against a wide range of bacteria); and

  • narrow spectrum (active against certain types of bacteria only, usually prescribed when a culture of a urine sample has identified a particular bacterium shown to be sensitive to the antibiotic, or when only a limited range of bacteria of known sensitivity are likely to be present).

In addition, the routes of administration (oral or intravenous, but not local or topical) were considered.

The following interventions were not considered in the present review:

  • catheterisation insertion techniques (e.g. clean, sterile, with or without antiseptic or antibiotic cream);

  • meatal care management techniques (e.g. routine hygiene, antiseptic or antibiotic cream); and

  • management of drainage systems (e.g. use of sterile or clean drainage bags, use of antiseptic solutions, washout/irrigation of drainage bags).

Types of outcome measures

Subjective and objective measures:
  • patient comfort;

  • patient satisfaction;

  • ease of use for patient;

  • sexual function;

  • incontinence/bypass leakage;

  • need to use supplementary pads/bed pads;

  • ease of use for practitioner;

  • need to change catheters;

  • number of catheters used;

  • length of time catheters used.

Complications/adverse effects:
  • asymptomatic bacteriuria;

  • symptomatic urinary tract infections;

  • use of prophylactic antibiotics;

  • use of rescue antibiotics;

  • urethral strictures;

  • bladder stones;

  • urgency/bladder spasms/detrusor overactivity;

  • other adverse effects of intervention (other than urinary tract infection).

Quality of life:
  • generic QoL measures (e.g. Short Form 36, Ware 1992);

  • psychological outcome measures (e.g. HADS, Zigmond 1983).

Economic outcomes:
  • costs of intervention(s);

  • resource implications of differences in outcomes;

  • formal economic analysis (cost-effectiveness, cost utility).

Any other non-prespecified outcomes judged to be important when performing the review.

Search methods for identification of studies

We did not impose any language restriction or other limits on the searches.

Electronic searches

This review has drawn on the search strategy developed for the Incontinence Group as a whole. Relevant trials have been primarily identified from the Cochrane Incontinence Group Specialised Trials Register. The methods used to derive this, including the search strategy, are described in the Group's module in The Cochrane Library. The register contains trials identified from MEDLINE, CINAHL and the Cochrane Central Register of Controlled Trials (CENTRAL) and handsearching of journals and conference proceedings.

The Incontinence Group Trials Register was searched using the Group's own keyword system. The search terms used were:

(design.rct* or design.cct*)
AND
({intvent.mech.cath*} or {intvent.mech.device*} or {intvent.mech.sheaths.} or {intvent.prevent.antibiotics*} or {intvent.prevent.antinfect.*} or {intvent.prevent.cath*} or {intvent.prevent.cleaning fluids*} or {intvent.prevent.surg*} or {intvent.surg.intraoperativemanagement*} or {intvent.surg.postsurgman*} or {intvent.surg.presurgman*.} or {intvent.surg.urethrotomy.})

(All searches were of the keyword field of Reference Manager 12, Thomson Reuters).
Date of the most recent search of the register for this review: 28 September 2011.
The trials in the Incontinence Group Specialised Register are also contained in CENTRAL.

Searching other resources

Additionally, all reference lists of identified trials were searched.

Data collection and analysis

Selection of studies

The reviewers independently assessed all titles and abstracts identified by the search. Where there was any possibility that the study could be included, the full paper was obtained. Any disagreement that could not be resolved by discussion was resolved by consultation with an independent third person.

Data extraction and management

Data were extracted by two reviewers independently and compared. If the data in trials were not fully reported, clarification was sought directly from the authors. Included trial data were processed as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005).

Assessment of risk of bias in included studies

Two reviewers independently assessed each report for inclusion and methodological quality using the Cochrane Incontinence Group's quality assessment tool: disagreements were resolved through discussion with the third reviewer. The Group's checklist covered quality of random allocation and concealment, description of dropout and withdrawals, analysis by intention-to-treat, and 'blinding' during treatment and at outcome assessment.

Measures of treatment effect

If the trialists did not pay attention to the time that patients were at risk, we calculated the incidence-density relative rates (IDR) and/or the incidence-density differences (IDD) and number needed to treat for an additional beneficial outcome (NNTB) or number needed to treat for an additional harmful outcome (NNTH) within a certain period (Bouter 1995). The incidence density was calculated using the following formula:

  • in the numerator: total number of events (bacteriuria or symptomatic urinary tract infection);

  • in the denominator: total person time of follow-up minus time not at risk (total number of events multiplied by the average number of days antibiotics were given plus total number of days receiving antibiotics for other reasons than event).

For categorical outcomes we related the numbers reporting an outcome to the numbers at risk in each group to derive a relative risk (RR) or to the total person time at risk (IDR). For continuous variables we used means and standard deviations to derive a weighted mean difference (WMD).

For cross-over trials, if data from the first arm of the trial were available, we analysed the data from the first period as if from a parallel-group trial and if the data had allowed, we would have combined results from both designs. If the data from the first arm of the trial were not available, the data were entered into Other Data Tables and comparisons were made only on the direction of any differences.

Unit of analysis issues

The unit of analysis was based on the individual patient (unit to be randomised for interventions to be compared).

Dealing with missing data

Irrespective of the type of data, we reported dropout rates in the 'Characteristics of included studies' table.

Assessment of heterogeneity

Trials were compared to assess and investigate the likelihood of important clinical heterogeneity. This also took into account the results of the Chi2 test for heterogeneity. As a general rule, the outcome data were combined using a fixed-effect model to calculate pooled estimates and their 95% confidence intervals. However, a random-effects model was considered where there were concerns that heterogeneity could be complicating an analysis.

Data synthesis

When appropriate, meta-analyses were undertaken. We performed statistical analyses according to the statistical guidelines referenced in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005).

Subgroup analysis and investigation of heterogeneity

Trials were subgrouped by whether the catheterisation was intermittent or indwelling.

Sensitivity analysis

If the data had allowed, we would have performed sensitivity analyses to assess the impact of trial quality, such as adequate versus poor allocation concealment but such data were not reported. Similarly, if the data had allowed, we would have performed subgroup analyses for different diagnostic groups or according to the sex of participants but such data were not reported.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies.

The literature search produced a total of 624 records which were screened for this review. The flow of literature through the assessment process is shown in the PRISMA flowchart ( Figure 1).

Figure 1.

PRISMA study flow diagram

Results of the search

Seven trials were included in the first version of the review (Anderson 1980; Duffy 1982 Gribble 1993 Johnson 1994 Mohler 1987a Rutschmann 1995 Schlager 1998). Four were cross-over trials (Duffy 1982; Johnson 1994;Rutschmann 1995;, Schlager 1998) and three were parallel-group randomised controlled trials (Anderson 1980; Gribble 1993 Mohler 1987a). One further trial was included in this update (Zegers 2011).

Included studies

Eight trials were included in this review (Anderson 1980; Duffy 1982; Gribble 1993; Johnson 1994; Mohler 1987a; Rutschmann 1995; Schlager 1998; Zegers 2011). Details are given in the Characteristics of included studies table.

Participants

Four trials involved adult patients with neurogenic bladder of differing types, all of whom were using intermittent catheterisation:

  • Two parallel-group trials included hospitalised adults with acute neurogenic bladder following recent spinal cord injury who were using intermittent self-catheterisation (Anderson 1980; Gribble 1993). In one of these trials, only male patients were included (Anderson 1980); in the other trial, almost seven times more males were included, but the distribution of men and women was equal between the trial groups (Gribble 1993).

  • One other parallel-group trial also involved adult hospitalised patients who were using intermittent catheterisation for neurogenic bladder, but it was not clear whether the neurogenic bladder was an acute problem or longstanding (Mohler 1987a). In that trial, three times more males were included, but the trial groups were comparable with regard to sex.

  • One cross-over trial included mainly male volunteers with neurogenic bladder using intermittent self-catheterisation at home (Duffy 1982); the reasons why patients suffered from neurogenic bladder were not reported.

  • One other cross-over trial included elderly nursing home inpatients with long-term urethral catheters (Rutschmann 1995). In that trial, indications for catheterisation were various.

Three trials involved children:

  • Two cross-over trials included children with neurogenic bladder due to meningomyelocele who were using intermittent self-catheterisation at home (Johnson 1994; Schlager 1998). In one of these trials, neurogenic bladder was due to spinal cord injury (Schlager 1998). Both trials involved more girls than boys.

  • In a randomised controlled trial with parallel groups only, children with neurogenic bladder due to spina bifida and clean intermittent catheterisation were studied, with more girls than boys (57% versus 43%) (Zegers 2011).

Duration of catheterisation

In the four cross-over trials, all patients were catheterised for longer than 14 days. In three trials, patients were catheterised for 12 weeks (Duffy 1982; Johnson 1994; Rutschmann 1995), and in one trial for 20 weeks (Schlager 1998). In the included parallel-group RCTs, the duration of catheterisation was not the same for all patients. In one trial, patients were catheterised for a mean of nine weeks, ranging from one to 16 weeks (Gribble 1993). In the other two trials, the longest patients were catheterised was until hospital discharge (Anderson 1980; Mohler 1987a); the mean duration was not reported for either of these trials. In one parallel-group RCT the follow-up was 18 months (Zegers 2011).

Types of intervention

All included trials, except two (Anderson 1980; Zegers 2011), randomised participants to prophylactic antibiotics or a matching placebo. Of the two remaining trials, in one (Anderson 1980) patients in the control group were openly allocated to no prophylaxis, and in the other trial (Zegers 2011) the participants were openly randomised to continuation or discontinuation of low dose prophylaxis.

Three different types of antibiotics were investigated:

With regard to the papers studying the effect of nitrofurantoin, patients in one trial (Duffy 1982) received twice as much nitrofurantoin as did patients in another trial (Anderson 1980). In two trials (Johnson 1994; Schlager 1998), similar approaches were used to adjust the dose of medication depending on children's weights. In the trials of TMX-SMX prophylaxis, the dose was four times higher in one (Mohler 1987a) than in the other (Gribble 1993). In one study, patients assigned to continuation of prophylactic antibiotic used various schemes in low doses, which included trimethoprim, nitrofurantoin, cefuroxime, co-trimoxazole or a combination of prophylactics (Zegers 2011).

Outcome measures

The definition of significant bacteriuria ranged from 102 colony forming units per ml to greater than 105 colony forming units per ml. The definition of symptomatic urinary tract infection did not always involve bacteriuria (Johnson 1994).

Two trials reported adverse events (Gribble 1993; Rutschmann 1995). In one (Gribble 1993), adverse events were defined as side effects due to trial medication; in the other (Rutschmann 1995), the definition incorporated both adverse events due to trial medication and catheter-related complications (i.e. obstruction, encrustation, leakage, suprapubic pain, inflammation of meatus, and haematuria).

To detect trial medication-resistant organisms, the Kirby-Bauer disk diffusion method was used in two trials (Gribble 1993; Rutschmann 1995).

Only one trial evaluated patients' general condition by the assessment of: 'well-being' of the patient; appreciation of his or her social involvement; awareness; need for nursing care; and functional status (Rutschmann 1995).

Excluded studies

Three studies were excluded (Kuhlemeier 1985; Maynard 1984; Warren 1982b). One did not address any of the questions posed in this review (Warren 1982b), and the others did not have sufficient information to judge the study methods (Kuhlemeier 1985) or extract outcome data (Maynard 1984). Two comparisons in one other study were also excluded (Mohler 1987b). One other trial (Clarke 2005) was excluded due to excessive losses after randomisations (38%) in unequal proportions between groups, which was considered to be high risk of attrition bias. The reasons for exclusion are listed in the table of Characteristics of excluded studies.

A further 14 studies were excluded for a variety of other reasons (see Characteristics of excluded studies).

Risk of bias in included studies

Details of the quality of each trial are given in the table of Characteristics of included studies.

Allocation

Not all trials described the group allocation method in detail. Two trials used an adequate method of concealment of allocation: in one trial (Gribble 1993), authors applied block randomisation in groups of eight which was pre-stratified using two factors: the presence or absence of prior bacteriuria; and the participant's sex. The randomisation sequence was obtained from a separate computer-generated randomisation schedule. A randomisation 'table' was used in one trial (Mohler 1987a). If the authors stated that a trial was randomised, double blind and placebo-controlled, we have assumed that allocation to groups was adequately concealed, and classed this as an adequate allocation concealment. Four trials were described as such (Duffy 1982; Johnson 1994; Rutschmann 1995; Schlager 1998). Only in one trial was the method of allocation concealment unclear (Anderson 1980). One study performed a computer based, random, concealed allocation, and only the outcome assessor was blind (Zegers 2011).

Blinding

The six trials with adequate randomisation were all placebo-controlled and hence clinical staff and patients were blind to the allocated intervention. Those testing urine samples were also blinded. Clinical staff and patients were not blinded in one trial (Anderson 1980): the participants were openly allocated antibiotics or no antibiotics. The authors did not report whether the outcome assessor was blinded (Anderson 1980).

Incomplete outcome data

In three trials, none of the participants were lost after randomisation (Anderson 1980; Duffy 1982; Zegers 2011), but in one, 38% of patients without prophylaxis switched to the prophylaxis group (Zegers 2011) as the authors performed a treatment effect analyses according to the intention-to-treat principle; this study was not excluded. In another two trials, loss after randomisation was 2% (Mohler 1987a) and 6% (Schlager 1998). In one trial (Gribble 1993), more than 30% of the randomised patients did not complete the trial. The authors reported the numbers of patients lost and reasons why patients were lost according to allocation to treatment. There were no indications of selective loss. In another trial (Rutschmann 1995) 35% of the randomised patients did not complete the trial and, in the last trial (Johnson 1994), 15% of randomised children were lost after randomisation. We were unable to judge whether or not the losses differed between the trial groups.

Selective reporting

We had no concerns about selective reporting since important clinical outcomes were evaluated.

Other potential sources of bias

Intention-to-treat analysis

Intention-to-treat analysis was followed in seven of the trials. In one trial (Anderson 1980), it was not clear whether this principle was followed.

Cross-over trials

With regard to the four cross-over trials, there was a one-month washout period in one of these (Schlager 1998); there was no washout period in the other three (Duffy 1982; Johnson 1994; Rutschmann 1995).

Effects of interventions

Comparison 1: Indwelling urethral catheterisation compared with suprapubic catheterisation

No eligible trials were found.

Comparison 2: Indwelling urethral catheterisation compared with intermittent catheterisation

No eligible trials were found.

Comparison 3: Suprapubic catheterisation compared with intermittent catheterisation

No eligible trials were found.

Comparison 4: Antibiotic prophylaxis compared with giving antibiotics when clinically indicated

Four eligible trials addressed this comparison (Johnson 1994; Rutschmann 1995; Schlager 1998; Zegers 2011). Three were cross-over trials and one was a parallel group controlled trial (Zegers 2011). Two trials tested nitrofurantoin against placebo in children using intermittent catheterisation (Johnson 1994; Schlager 1998); one trial tested norfloxacin against placebo in elderly patients using indwelling urethral catheterisation (Rutschmann 1995). One trial enrolled children using intermittent catheterisation who were in use of different regimens of prophylaxis and randomised to continuation or discontinuation with follow-up of 18 months and outcome assessor blinded (Zegers 2011).

Bacteriuria (asymptomatic and symptomatic)

Bacteriuria was not considered as an outcome measure in this review because asymptomatic bacteriuria was not treated and, therefore, it was not a useful measure in this setting.

Symptomatic urinary tract infection (UTI)
a) Participants with trials of intermittent catheterisation

For two trials we calculated the rate of symptomatic UTI per catheterisation week (Johnson 1994; Schlager 1998). In a trial involving 15 children, those receiving prophylactic antibiotics ran almost half of the risk for UTI, but this was not statistically significant (IDR 0.50, 95% CI 0.17 to 1.44, Analysis 4.1.1 (Schlager 1998). Of the 15 participants, eight had at least one UTI while taking prophylactic antibiotics compared with 11 when taking placebo (Analysis 4.3.1) (Schlager 1998). In the bigger trial (Johnson 1994), there were four cases of UTI in 430 catheterised weeks during antibiotic prophylaxis compared with two cases in 389 control weeks without antibiotics (Analysis 4.2.1) (Johnson 1994). In spite of being clinically and methodologically homogeneous, the two trials thus showed inconsistent results for symptomatic UTI.

In the third trial (Zegers 2011), those who continued taking prophylaxis had fewer episodes of at least one febrile symptomatic UTI, but this was not statistically significant (RR 0.50, 95% CI 0.09 to 2.66) (Analysis 4.10.1) ). However, the antibiotic prophylaxis group did have significantly fewer afebrile symptomatic UTIs (IDR 0.69, 95% CI 0.55 to 0.87) (Analysis 4.11.1), with number needed to treat for an additional harmful outcome of 2.2.

It was not possible to perform meta-analysis of the three trials due heterogeneity in the method of classification of UTI and outcome measures.

b) Trials of indwelling urethral catheterisation

In the open trial of prophylaxis amongst patients with an indwelling catheter, there were fewer symptomatic UTIs during antibiotic prophylaxis than during control periods without antibiotics (one in 276 catheterisation weeks versus 12 in 259 weeks) (Analysis 4.2.2)) (Rutschmann 1995).

Other outcome measures
a) Participants with intermittent catheterisation

No data on other outcomes were reported.

b) Participants with indwelling urethral catheterisation

Other outcome measures were reported only for the single trial of patients with indwelling urethral catheters. In this trial, the results in the antibiotic group were better in terms of:

However, for the outcome 'percentage of resistant strains', the results were in favour of the control group (Analysis 4.8) (Rutschmann 1995).

Comparison 5: Antibiotic prophylaxis compared with giving antibiotics when microbiologically indicated

Four eligible trials were found (Anderson 1980; Duffy 1982; Gribble 1993; Mohler 1987a). Three were parallel-group randomised controlled trials (Anderson 1980; Gribble 1993; Mohler 1987a) and one was a cross-over trial (Duffy 1982). The trial populations were all adult patients using intermittent catheterisation. One trial (Anderson 1980) tested nitrofurantoin against no nitrofurantoin; another trial (Duffy 1982) tested nitrofurantoin against placebo and two trials tested TMP-SMX against placebo (Gribble 1993; Mohler 1987a).

Asymptomatic bacteriuria
a) Participants with intermittent catheterisation

All four trials provided data on bacteriuria. For three trials we were able to calculate the rate of bacteriuria per catheterisation week (Anderson 1980; Duffy 1982; Mohler 1987a). In one trial (Gribble 1993), this was not possible; the authors of that trial presented the outcome as having at least one episode of bacteriuria.

The two parallel-group trials found that patients in the prophylactic antibiotic group had fewer episodes of bacteriuria than patients not receiving prophylactic antibiotics (IDR of the pooled analysis 0.61, 95% CI 0.44 to 0.87, Analysis 5.1.1; IDD -0.14, 95% CI -0.23 to -0.05, Analysis 5.3) (Anderson 1980; Mohler 1987a) although there was evidence of heterogeneity. The results of the cross-over trial were also in favour of the prophylaxis group (Analysis 5.2) (Duffy 1982). Anderson (Anderson 1980) also analysed the effect according to the frequency of catheterisation and found some evidence that the effect of prophylaxis was greater in the subgroup who used intermittent catheterisation every four hours rather than every eight hours (every four hours: IDR 0.15, 95% CI 0.05 to 0.42, Analysis 5.1.2; every eight hours: IDR 0.49, 95% CI 0.21 to 1.12, Analysis 5.1.3) (Anderson 1980).

Gribble (Gribble 1993) found that although there were fewer patients having at least one episode of bacteriuria in the antibiotic group, this did not quite reach statistical significance (RR 0.86, 95% CI 0.72 to 1.02, Analysis 5.7.1) (Gribble 1993). The authors reported the effect separately for men and women and found similar RRs (for males: RR 0.85, 95% CI 0.71 to 1.03, Analysis 5.7.2; for females: RR 0.89, 95% CI 0.57 to 1.38, Analysis 5.7.3) (Gribble 1993).

b) Participants with indwelling urethral catheterisation

No eligible trials were found.

Symptomatic bacteriuria
a) Participants with intermittent catheterisation

Two trials (Gribble 1993; Mohler 1987a) reported data on symptomatic bacteriuria. For one trial (Mohler 1987a) we calculated the rate of symptomatic bacteriuria per catheterisation week. The results showed that patients in the prophylactic antibiotic group had a lower rate than patients not receiving prophylactic antibiotics but this was not statistically significant (IDR 0.56, 95% CI 0.27 to 1.15, Analysis 5.5) (Mohler 1987a).

In the other trial, fewer patients in the prophylactic antibiotic group had at least one episode of symptomatic bacteriuria, and this was statistically significant (RR 0.19, 95% CI 0.07 to 0.53, Analysis 5.8.1; RD -0.26, 95% CI -0.39 to -0.13, Analysis 5.9) (Gribble 1993). Most participants were male and all except one of the episodes was in a man (Analysis 5.8.2 and Analysis 5.8.3) (Gribble 1993).

b) Participants with indwelling urethral catheterisation

No eligible trials were found.

Other outcome measures
a) Participants with intermittent catheterisation

Other outcome measures were reported for only one trial (Gribble 1993). We calculated the rate of adverse events per catheterisation week. The results were better in the antibiotic group but were not statistically significant (IDR 0.74, 95% CI 0.53 to 1.02, Analysis 5.6) (Gribble 1993). There was also no statistically significant difference in the number of patients having at least one episode of adverse events (RR 0.86, 95% CI 0.64 to 1.14, Analysis 5.10) (Gribble 1993). Fewer patients in the prophylaxis group were prescribed antibiotics for at least one episode of urinary tract infection (RR 0.78, 95% CI 0.62 to 0.97, Analysis 5.11) (Gribble 1993). There were also few data about possible antibiotic resistance. Although there were no differences in: at least one episode of bacteriuria due to TMX-SMX-resistant organisms (RR 0.95, 95% CI 0.77 to 1.17, Analysis 5.12) (Gribble 1993); and, at least one time recovery of TMX-SMX-resistant gram-negative bacilli from weekly surveillance cultures (RR 1.17, 95% CI 0.80 to 1.72, Analysis 5.13) (Gribble 1993), the confidence intervals were all wide.

b) Participants with indwelling urethral catheterisation

No eligible trials were found.

Comparison 6: Giving antibiotics, if microbiologically indicated, compared with giving antibiotics if clinically indicated

No eligible trials were found.

Discussion

Summary of main results

This systematic review identified eight eligible trials that addressed two of the six pre-stated hypotheses. Only one of these was amongst frail, elderly patients; the others studied patients (adults or children) with neurogenic bladder problems using intermittent catheterisation. We found no trials that compared indwelling urethral catheterisation with suprapubic catheterisation, indwelling urethral catheterisation with intermittent catheterisation, suprapubic catheterisation with intermittent catheterisation, or giving antibiotics if microbiologically indicated with giving antibiotics if clinically indicated (Comparisons 1 to 3, or 6). Six out of the eight included trials of different antibiotic policies included patients using intermittent catheterisation, despite there being no RCT that establishes that intermittent catheterisation is the preferable policy. It would be clinically valuable to conduct trials to address the question whether one technique is superior to another.

We analysed whether antibiotic prophylaxis was better than giving antibiotics when clinically indicated (comparison 4); this included three cross-over trials that analysed a total of 94 catheterised patients (71 with intermittent catheterisation, 23 with indwelling urethral catheterisation) and one parallel group trial with 176 participants with intermittent catheterisation. Several types of antibiotics were tested: nitrofurantoin, norfloxacin, trimethoprim, cefuroxime, co-trimoxazole or a combination of prophylactics. All four trials used an adequate method of randomisation concealment and all four provided the outcome measure symptomatic UTI. Only one of the trials addressed other outcome measures, such as visual encrustation, catheter obstructions, adverse events, general condition, percentage of resistant strains and spectrum of microbial isolates. Therefore, there were few useful data for addressing concerns about possible adverse effects of antibiotic prophylaxis.

The results of the patients using intermittent catheterisation were kept separate from those of patients with indwelling urethral catheterisation; the pathway of getting bacteriuria is different and it is conceivable, therefore, that patients run different risks of infection. It is clinically important to know whether the effect of prophylactic antibiotics is concentrated in certain groups.

For patients using intermittent catheterisation, there was some inconsistency between the trials. In the two cross-over studies, there were no significant differences between groups, with fewer UTIs in the intervention group in one study (Schlager 1998), and fewer UTIs in the control group in another study (Johnson 1994). The additional trial included in the update of this review (Zegers 2011) separated the classification of symptomatic UTI into febrile and afebrile, a method not used in previous studies. For febrile UTI there was no significant difference between the group with prophylactic antibiotics and control. However, afebrile UTI occurred significantly less frequently in the group with prophylactic antibiotics; the number needed to treat for an additional harmful outcome was 2.2.

For patients using indwelling urethral catheterisation, the one small trial suggested that prophylaxis reduced symptomatic UTI. This trial also favoured the prophylactic antibiotic group in respect of other outcomes, except for microbial resistance pattern. As expected, the trial suggested that an important disadvantageous effect of prophylactic antibiotics was the selection of antimicrobial-resistant micro-organisms.

We did not include bacteriuria (asymptomatic and symptomatic) as an outcome measure, because patients who had previously had untreated asymptomatic bacteriuria would run a greater risk of having a repeated positive urine culture. Thus, there was a strong interdependence of the various contributions of bacteriuria, which might have led to invalid results.

We also analysed if antibiotic prophylaxis is better than giving antibiotics when microbiologically indicated (comparison 5), and included three RCTs and one cross-over trial, involving a total of 234 patients using intermittent catheterisation. There were two types of antibiotics tested: nitrofurantoin and TMX-SMX. Three trials used an adequate method of randomisation concealment, one trial an inadequate method. All four trials provided the outcome measure bacteriuria (asymptomatic and symptomatic). Two trials addressed the outcome measure symptomatic bacteriuria and a single trial included the outcome measures adverse events, need of antibiotics for UTI, bacteriuria due to TMX-SMX-resistant organisms and recovery of TMX-SMX-resistant gram-negative bacilli.

The limited evidence available suggested that receiving prophylactic antibiotics reduced the rate of bacteriuria (asymptomatic and symptomatic). The results of the pooled incidence rate difference in bacteriuria (asymptomatic and symptomatic) suggested about 14 fewer periods of bacteriuria within two years when patients received prophylactic antibiotics (IDD -0.14, 95% CI -0.23 to -0.05, Analysis 5.3). There was weak evidence that prophylactic antibiotics were better in terms of less symptomatic bacteriuria and less need for antibiotics for UTIs. There were no statistically significant differences in the effects on adverse events, bacteriuria due to TMX-SMX-resistant organisms and recovery of TMX-SMX-resistant gram-negative bacilli from weekly surveillance cultures, but confidence intervals were wide.

We could not calculate the risk difference in symptomatic bacteriuria for Gribble's study, because the follow-up of patients was so variable (mean time in the trial was nine weeks in both groups, with a range from one to 16 weeks) (Gribble 1993).

Quality of the evidence

The quality of the trials and their reporting were generally unsatisfactory. All the included trials had small sample sizes and overall 504 patients were randomised. Therefore, estimates of differential effects were imprecise, with wide confidence intervals. Just two trials clearly had adequate concealment of randomisation although it was assumed in five others. Two trials did not report whether there were dropouts and in another two trials more than 30% of the randomised participants were lost. It was frustrating that the data from the cross-over trials were reported as if they were from a parallel group trial. As a results, we could not analyse them in MetaView as the appropriate statistics were not available. Given the nature of the intervention, the cross-over trials might have been expected to have a 'washout' period, but this occurred in only one trial. The possibilities for combining data were also limited by clinical heterogeneity.

Potential biases in the review process

The choice of measure of subsequent UTI is not straightforward. Most of the trials did not pay attention to the fact that, in the long-term catheterisation setting, trial participants could be infected more than once. Patients who had had bacteriuria or a clinical UTI were treated with systemic antibiotics. During a period of having bacteriuria or a clinical symptomatic UTI and receiving systemic antibiotics, patients were not at risk for the intended outcome measure. In addition, within the parallel-group RCTs, patients were followed over differing periods of time. We have attempted to obtain more valid results by transforming the data of most of the included trials using the methods of incidence density. However, this approach does not take into account possible 'clustering' of events in a few individuals; for this reason we have also reported (where available) the number of participants who ever had an infection.

Another concern is that it is conceivable that the effect of prophylactic antibiotics is different in patients learning the technique of intermittent catheterisation from patients who had mastered the technique. None of the trialists controlled for that in the analysis.

In summary, the clinically most relevant measure for infection is symptomatic UTI. The effect of prophylactic antibiotics on symptomatic UTI in the long-term setting can be measured by the rate of infection or by the number of patients having at least one infection. The rate of infection allows us to detect whether patients allocated to the prophylaxis group had fewer infections within a certain period. The number of patients having at least one episode of infection enables us to detect whether clinical infections are limited to certain patients. Future trials should pay more attention to these considerations.

Agreements and disagreements with other studies or reviews

Despite the relevance of the subject, so far, few studies of quality have evaluated the effectiveness and safety of various urinary catheter policies for long-term bladder drainage. This lack of evidence was also found in other systematic reviews. One that aimed to analyse the catheter policies for management of long-term voiding problems in adults with neurogenic bladder disorders, concluded that there have not been any randomised trials to provide good evidence (Jamison 2011). Another review evaluated the long-term bladder management by intermittent catheterisation in adults and children, and concluded that it is not possible to state that one catheter type, technique or strategy is better than another due to insufficient quality trials (Moore 2007). Finally, a systematic review about types of indwelling urinary catheters for long-term bladder drainage in adults concluded that the evidence was not sufficient as a reliable basis for practical conclusions because all trials included were small and showed methodical weaknesses (Jahn 2007).

Authors' conclusions

Implications for practice

Is indwelling urethral catheterisation better than suprapubic catheterisation?

No eligible trials were identified that addressed this question.

Is indwelling urethral catheterisation better than intermittent catheterisation?

No eligible trials were identified that addressed this question.

Is suprapubic catheterisation better than intermittent catheterisation?

No eligible trials were identified that addressed this question.

Is antibiotic prophylaxis better than giving antibiotics when clinically indicated?

The evidence available is not sufficient as a basis for determining practice. For patients using intermittent catheterisation the data were inconclusive with only one of three studies with statistically significant differences favouring prophylaxis. For patients using indwelling urethral catheterisation, only a single inconclusive cross-over trial investigated this issue.

Is antibiotic prophylaxis better than giving antibiotics when microbiologically indicated?

For patients using intermittent catheterisation, the limited evidence available suggested that antibiotic prophylaxis reduces the number of episodes of bacteriuria (asymptomatic and symptomatic). For patients using urethral catheterisation, limited data were available.

Is giving antibiotics, if microbiologically indicated, better than giving antibiotics if clinically indicated?

None of the trials included in the review address this question.

Implications for research

There is a need for large randomised trials to determine the optimal catheter policies for the long-term bladder drainage in adults and children. The most important patient groups are the frail, elderly and younger people with neurogenic bladder problems. Although there was some evidence in favour of antibiotic prophylaxis, the implications of general adoption of this policy are unclear and should be evaluated especially in terms of adverse effects, development of antibiotic resistance and economic implications. Trials comparing intermittent, indwelling urethral and suprapubic policies are also needed. Cross-over trials should report their data correctly, i.e. the difference between the treatments and its standard error.

Primary outcomes should include symptomatic urinary tract infection, measured both by the rate of infection and the number of patients having at least one episode of infection, and the development of resistant strains, preferably investigated over a longer period than the trials to date. With respect to patients receiving intermittent catheterisation, the outcome infection should be reported separately for patients learning the technique as opposed to those who have mastered the technique.

Acknowledgements

We would like to thank Sheila Wallace from the Cochrane Incontinence Group for her collaboration.

Data and analyses

Download statistical data

Comparison 4. Prophylactic antibiotic versus antibiotic for symptomatic UTI
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Rate of symptomatic UTI per catheterisation week - cross-over trial1 IDR, 95% CI (Fixed, 95% CI)Totals not selected
1.1 Intermittent catheterisation1 IDR, 95% CI (Fixed, 95% CI)0.0 [0.0, 0.0]
2 Rate of symptomatic UTI - cross-over trial  Other dataNo numeric data
2.1 Intermittent catheterisation (both periods)  Other dataNo numeric data
2.2 Urethral catheterisation (both periods)  Other dataNo numeric data
3 At least one episode of symptomatic urinary tract infection - cross-over trial  Other dataNo numeric data
3.1 Intermittent catheterisation (both periods)  Other dataNo numeric data
4 Rate of visual encrustation - cross-over trial  Other dataNo numeric data
4.1 Urethral catheterisation (both periods)  Other dataNo numeric data
5 Rate of catheter obstruction - cross-over trial  Other dataNo numeric data
5.1 Urethral catheterisation (both periods)  Other dataNo numeric data
6 Rate of adverse events - cross-over trial  Other dataNo numeric data
6.1 Urethral catheterisation (both periods)  Other dataNo numeric data
7 Patients general condition - cross-over trial  Other dataNo numeric data
7.1 Urethral catheterisation (both periods)  Other dataNo numeric data
8 Microbial resistance pattern (resistant strains/total number of strains) - cross-over trial  Other dataNo numeric data
8.1 Urethral catheterisation (both periods)  Other dataNo numeric data
9 Spectrum of microbial isolates (number of gram-neg. strains/total number of strains - cross-over trial  Other dataNo numeric data
9.1 Urethral catheterisation (both periods)  Other dataNo numeric data
10 At least one episode of febrile symptomatic UTI. Parallel group trial1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
10.1 Intermittent catheterisation1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
11 Rate of afebrile symptomatic UTI per person year. Parallel group trial1 IDR, 95% CI (Fixed, 95% CI)Totals not selected
11.1 Intermittent catheterisation1 IDR, 95% CI (Fixed, 95% CI)0.0 [0.0, 0.0]
Analysis 4.1.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 1 Rate of symptomatic UTI per catheterisation week - cross-over trial.

Analysis 4.2.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 2 Rate of symptomatic UTI - cross-over trial.

Rate of symptomatic UTI - cross-over trial
StudyTreatmentControl
Intermittent catheterisation (both periods)
Johnson 19944 UTIs in 430 catherisation weeks2 UTIs in 389 catheterisation weeks
Urethral catheterisation (both periods)
Rutschmann 19951 UTI in 276 catheterisation weeks12 UTIs in 259 catheterisation weeks

Analysis 4.3.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 3 At least one episode of symptomatic urinary tract infection - cross-over trial.

At least one episode of symptomatic urinary tract infection - cross-over trial
StudyTreatmentControl
Intermittent catheterisation (both periods)
Schlager 19988/15 study participants11/15 study participants

Analysis 4.4.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 4 Rate of visual encrustation - cross-over trial.

Rate of visual encrustation - cross-over trial
StudyTreatmentControl
Urethral catheterisation (both periods)
Rutschmann 19954 events in 276 catheterisation weeks19 events in 259 catheterisation weeks

Analysis 4.5.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 5 Rate of catheter obstruction - cross-over trial.

Rate of catheter obstruction - cross-over trial
StudyTreatmentControl
Urethral catheterisation (both periods)
Rutschmann 19952 events in 276 catheterisation weeks8 events in 259 catheterisation weeks

Analysis 4.6.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 6 Rate of adverse events - cross-over trial.

Rate of adverse events - cross-over trial
StudyTreatmentControl
Urethral catheterisation (both periods)
Rutschmann 1995596 events in 276 catheterisation weeks744 events in 259 catheterisation weeks

Analysis 4.7.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 7 Patients general condition - cross-over trial.

Patients general condition - cross-over trial
StudyTreatmentControl
Urethral catheterisation (both periods)
Rutschmann 199512 patients improved / 23 study participants1 patient improved / 23 study participants

Analysis 4.8.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 8 Microbial resistance pattern (resistant strains/total number of strains) - cross-over trial.

Microbial resistance pattern (resistant strains/total number of strains) - cross-over trial
StudyTreatmentControl
Urethral catheterisation (both periods)
Rutschmann 199520 resistant strains/22 isolated strains8 resistant strains/41 isolated strains

Analysis 4.9.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 9 Spectrum of microbial isolates (number of gram-neg. strains/total number of strains - cross-over trial.

Spectrum of microbial isolates (number of gram-neg. strains/total number of strains - cross-over trial
StudyTreatmentControl
Urethral catheterisation (both periods)
Rutschmann 19955 gram-neg. strains/ 22 isolated strains31 gram-neg. strains/ 41isolated strains
Analysis 4.10.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 10 At least one episode of febrile symptomatic UTI. Parallel group trial.

Analysis 4.11.

Comparison 4 Prophylactic antibiotic versus antibiotic for symptomatic UTI, Outcome 11 Rate of afebrile symptomatic UTI per person year. Parallel group trial.

Comparison 5. Prophylactic antibiotic versus antibiotic for laboratory UTI
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Rate of bacteriuria (asymptomatic and symptomatic) per catheterisation week2 IDR, 95% CI (Fixed, 95% CI)Subtotals only
1.1 Overall277IDR, 95% CI (Fixed, 95% CI)0.61 [0.44, 0.87]
1.2 Sub-category: catheterisation every four hours12IDR, 95% CI (Fixed, 95% CI)0.15 [0.05, 0.42]
1.3 Sub-category: catheterisation every eight hours12IDR, 95% CI (Fixed, 95% CI)0.49 [0.21, 1.12]
2 Rate of bacteriuria (asymptomatic and symptomatic) - cross-over trial (both periods)  Other dataNo numeric data
3 Rate difference of bacteriuria (asymptomatic and symptomatic) per catheterisation week2 IDD (Fixed, 95% CI)Subtotals only
3.1 Overall277IDD (Fixed, 95% CI)-0.14 [-0.23, -0.05]
4 Rate difference of bacteriuria per catheterisation week - cross-over trial1 IDD (Fixed, 95% CI)Totals not selected
5 Rate of symptomatic bacteriuria per catheterisation week1 IDR, 95% CI (Fixed, 95% CI)Totals not selected
6 Rate of adverse events per catheterisation week1 IDR (Fixed, 95% CI)Totals not selected
7 At least one episode of bacteriuria (asymptomatic and symptomatic)1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
7.1 Overall1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
7.2 Sub-category: male1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
7.3 Sub-category: female1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
8 At least one episode of definite symptomatic bacteriuria1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
8.1 Overall1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
8.2 Sub-category: male1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
8.3 Sub-category: female1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
9 At least one episode of definite symptomatic bacteriuria1 Risk Difference (M-H, Fixed, 95% CI)Totals not selected
10 At least one episode of adverse events (subject with any event)1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
11 At least one episode of need of antibiotics for urinary tract infection1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
12 At least one episode of bacteriuria due to TMX-SMX-resistant organisms1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
13 At least one time recovery of TMX-SMX-resistant gram-negative bacilli from weekly surveillance cultures1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
Analysis 5.1.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 1 Rate of bacteriuria (asymptomatic and symptomatic) per catheterisation week.

Analysis 5.2.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 2 Rate of bacteriuria (asymptomatic and symptomatic) - cross-over trial (both periods).

Rate of bacteriuria (asymptomatic and symptomatic) - cross-over trial (both periods)
StudyTreatmentControl
Duffy 19829 events in 90 weeks25 events in 85 weeks
Analysis 5.3.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 3 Rate difference of bacteriuria (asymptomatic and symptomatic) per catheterisation week.

Analysis 5.4.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 4 Rate difference of bacteriuria per catheterisation week - cross-over trial.

Analysis 5.5.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 5 Rate of symptomatic bacteriuria per catheterisation week.

Analysis 5.6.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 6 Rate of adverse events per catheterisation week.

Analysis 5.7.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 7 At least one episode of bacteriuria (asymptomatic and symptomatic).

Analysis 5.8.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 8 At least one episode of definite symptomatic bacteriuria.

Analysis 5.9.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 9 At least one episode of definite symptomatic bacteriuria.

Analysis 5.10.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 10 At least one episode of adverse events (subject with any event).

Analysis 5.11.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 11 At least one episode of need of antibiotics for urinary tract infection.

Analysis 5.12.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 12 At least one episode of bacteriuria due to TMX-SMX-resistant organisms.

Analysis 5.13.

Comparison 5 Prophylactic antibiotic versus antibiotic for laboratory UTI, Outcome 13 At least one time recovery of TMX-SMX-resistant gram-negative bacilli from weekly surveillance cultures.

What's new

DateEventDescription
9 July 2012New citation required but conclusions have not changedTwo new studies identified and one included in the review.
9 July 2012New search has been performedTwo new studies identified and one included in the review.

History

Protocol first published: Issue 2, 2003
Review first published: Issue 1, 2005

DateEventDescription
3 August 2009New search has been performedMinor update, no new studies. Twelve new studies identified from Incontinence register; all were excluded.
13 October 2008AmendedConverted to new review format.
10 May 2006New search has been performedfirst update
17 November 2004New citation required and conclusions have changedSubstantive amendment

Contributions of authors

All reviewers contributed to writing the review.

LAS and EMKS updated the review.

Declarations of interest

None known.

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • National Health Service Research and Development Programme, UK.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Anderson 1980

MethodsRCT
Participants

Inclusion criteria: 31 hospitalised male patients with spinal cord injury and acute neurogenic bladder (within 30 days of injury) starting with sterile intermittent catheterisation every 4 or 8 hours to keep the bladder volume < 500 ml

Exclusion criteria: No antibiotic prophylaxis

Interventions

Prophylaxis (15):
100 mg nitrofurantoin macrocrystals prophylaxis once a day (oral)

Control (16):
No antibiotics

Duration of intervention:
1) Until a balanced bladder status was achieved (defined as post-void residual volume < 100 ml);
2) Hospital discharge

OutcomesRate of bacteriuria per catheterisation day (symptomatic and asymptomatic)
Overall:
P:18/955; C: 36/686
In subgroups:
a) Catheterisation frequency every 4 hours: P: 4/577; C: 27/568
b) Catheterisation frequency every 8 hours: P: 14/378; C: 9/118
NotesSignificant bacteriuria defined as 10,000 cfu/ml in fresh catheterised specimen
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomisation table
Allocation concealment (selection bias)Low riskAdequate
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo losses
Selective reporting (reporting bias)Low riskAll outcomes related

Duffy 1982

MethodsCross-over study
ParticipantsInclusion criteria: 31 volunteers with neurogenic bladder using intermittent self-
catheterisation every 4 hours
Interventions

Prophylaxis (31):
100 mg nitrofurantoin macrocrystals prophylaxis twice daily for three months (oral)

Control (31):
Matching placebo (oral) twice daily for three months

Duration of intervention:
six months

OutcomesRate of bacteriuria per catheterisation week
(symptomatic + asymptomatic)
P: 9/90; C: 25/85
NotesSignificant bacteriuria defined as > 100,000 cfu/ml of at least one organism > 100,000 cfu/ml
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot reported
Allocation concealment (selection bias)Unclear riskNot reported
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blind
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo losses
Selective reporting (reporting bias)Low riskAll outcomes related

Gribble 1993

MethodsRCT
Participants

Inclusion criteria: 131 hospitalised patients with
1) Acute spinal cord injury 30 days or less prior to enrolment
2) Neurogenic bladder dysfunction
3) Commencement of clean intermittent catheterisation 72 hours or less prior to enrolment
4) Age 18 years or older

Exclusion criteria:
1) Known hypersensitivity to TMP or sulfenamide
2) Serum creatinine greater than 176 µmol/L
3) Serum transaminase levels greater than 3 times normal
4) Pregnancy

Interventions

Prophylaxis (66):
Low-dose TMP-SMX once daily (TMP 40 mg, SMX 200 mg) (oral)

Control (60):
Matching placebo

Duration of intervention
1) Until intermittent catheterisation was discontinued
2) Until discharge
3) For a maximum of 16 weeks

Outcomes

At least one episode of bacteriuria (symptomatic + asymptomatic):
Overall:
P: 49/66; C: 52/60
In subgroups:
a) In males: P: 42/57; C: 45/52
b) In females: P: 7/9; C: 7/8

At least 1 episode of definite symptomatic bacteriuria:
Overall:
P: 4/66; C: 19/60
In subgroups:
a) In males: P: 4/57; C: 18/52
b) In females: P: 0/9; C; 1/8

At least 1 adverse event:
P: 37/67; C: 40/62

At least 1 episode of need of antibiotics for UTI:
P: 41/66; C: 48/60

At least 1 episode of bacteriuria due to TMX-SMX resistant organisms:
P: 47/66; C: 45/60

At least 1 time recovery of TMX-SMX resistant gram-negative bacilli from weekly surveillance cultures:
P: 33/67; C: 26/62

Rate of adverse events per catheterisation week:
P: 66/613; C: 85/582

Notes

Significant bacteriuria defined as ≥ 100 cfu/ml of an organism in 2 consecutive urine specimens or ≥
100,000 from a single specimen

Symptomatic urinary tract infection defined as having bacteriuria and fever > 38 °C and at least one classical manifestation of UTI

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskBlock randomisation in groups of eight which was pre-stratified into two groups: presence or absence of prior bacteriuria and sex; the randomisation scheme was obtained from a separate computer-generated randomisation schedule
Allocation concealment (selection bias)Low riskAdequate
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blind
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind
Incomplete outcome data (attrition bias)
All outcomes
High riska) Excluded from study: Control: 2 did not receive study medication
b) 44 patients did not complete the study:
Prophylaxis: 12 second relapse; 7 suspected adverse effects; 3 withdrew consent; 3 others. Control: 10 second relapse; 5 suspected adverse effects; 2 withdrew consent; 2 others
c) 3 patients were excluded from efficacy analysis: prophylaxis: 1; control: 2
Selective reporting (reporting bias)Low riskAll outcomes related

Johnson 1994

MethodsCross-over study (method of allocation not stated, double-blind)
ParticipantsInclusion criteria: 66 children with neurogenic bladder due to meningomyelocele and clean intermittent catheterisation at home
Interventions

Prophylaxis (56):
Nitrofurantoin once daily (oral): 25 mg nitrofurantoin for children weighing between 12.5 and 25 kg; 50 mg nitrofurantoin for children > 25 kg

Control (56):
Matching placebo

Duration of intervention:
24 weeks (2 times 12 weeks)

OutcomesRate of clinical urinary tract infection per catheterisation week:
P: 4/430; C: 2/389
NotesClinical infection defined as one or more of the following: flank or abdominal pain, pyrexia, incontinence ascribed by a physician to a urinary tract infection
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot stated
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blind
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind
Incomplete outcome data (attrition bias)
All outcomes
Low riska) Excluded from study: 10 patients were not started on study medication
b) 4 patients dropped out but were included in the intention- to- treat analysis: 1 patient in group I during the first period (antibiotic prophylaxis) due to an increased number of UTIs; 3 patients in group II during the first period (placebo): 2 patients because of dissatisfaction with the number of UTIs and 1 family moved
Selective reporting (reporting bias)Low riskAll outcomes related

Mohler 1987a

MethodsRCT (randomisation table, double-blind)
ParticipantsInclusion criteria: 47 adult patients admitted to a rehabilitation facility with sterile intermittently catheterised neurogenic bladder
Interventions

Prophylaxis (21):
160 mg trimethoprim and 800 mg sulfamethoxazole once a day (oral)

Control (25):
Matching placebo

Duration of intervention:
until hospital discharge

Outcomes

Rate of bacteriuria per catheterisation day (symptomatic + asymptomatic):
P: 36/914; C: 48/1022

Rate of asymptomatic bacteriuria:
P: 25/914; C: 26/1022

Rate of symptomatic bacteriuria:
P: 11/914; C: 22/1022

Notes

Asymptomatic urinary tract infection defined as
a) ≥ 10,000 cfu/ml of a single or predominant species in a (sterile) catheterised specimen
b) ≥ 10,000 cfu/ml of a single or predominant species in a clean-catch specimen in male patients with suspected symptomatic UTI
c) ≥ 100,000 cfu/ml of a single or predominant species in a clean-catch specimen in asymptomatic male patients

Symptomatic urinary tract infection defined as asymptomatic UTI plus T > 38 °C , white blood cell count > 104, clinical symptoms of UTI and no other apparent source of infection

Total UTI defined as symptomatic and asymptomatic UTI

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomisation table
Allocation concealment (selection bias)Low riskAdequate
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blind
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind
Incomplete outcome data (attrition bias)
All outcomes
Low risk1 patient was eliminated from the study because of protocol violation (extended antibiotic administration)
Selective reporting (reporting bias)Low riskAll outcomes related

Rutschmann 1995

MethodsCross-over study
Participants

Inclusion criteria: 34 elderly inpatients (nursing home) with long-term urethral catheters

The results of 23 patients were analysed

Interventions

Prophylaxis (34):
200 mg norfloxacin daily for 3 months (oral)

Control (34):
Matching placebo

Duration of intervention:
6 months

Outcomes

Rate of symptomatic UTI per catheterisation week:
P: 1/276; C: 12/259

Rate of visual encrustation per catheterisation week:
P: 4/276; C: 19/259

Rate of catheter obstruction per catheterisation week:
P: 2/276; C: 8/259

Rate of adverse events per catheterisation week:
P: 596/276; C: 744/259

Patients' general condition (number of patients not improved/ number of study participants):
P: 12/23; C: 1/23

Microbial resistance pattern (resistant strains/total no. of strains)
P: 20/22; C: 8/41

Spectrum of microbial isolates (no. of gram-negative strains/total no. of strains)
P: 5/22; C: 31/41

Notes

Symptomatic urinary tract infection defined as bacteriuria ≥ 100,000 cfu/ml and

a) T ≥ 38.5 °C for two consecutive days in the absence of other clinical sources of infection or
b) Flank pain or unexplained mental disturbance or abdominal discomfort

Adverse events defined as obstruction, encrustation, leakage, suprapubic pain, inflammation of meatus, haematuria and side effects of treatment.

Patients' general condition defined as unchanged during the entire study period or improved during one or the other period. Criteria included well-being of the patient, appreciation of his or her social involvement, awareness, need for nursing care and functional status

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot stated
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blind
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind
Incomplete outcome data (attrition bias)
All outcomes
High risk11 patients did not complete the study: In 4 patients the catheter was withdrawn (numbers not reported according to allocated group); Prophylaxis: 3 died of non-infectious causes; Control: 3 died of non-infectious causes, 1 died of septic shock
Selective reporting (reporting bias)Low riskAll outcomes related

Schlager 1998

MethodsCross-over study
ParticipantsInclusion criteria: 16 children with chronic neurogenic bladder using clean intermittent catheterisation 4 times a day and living at home (14 meningomyelocele, 1 spinal cord injury)
Interventions

Prophylaxis (16):
Nitrofurantoin once daily (25 mg nitrofurantoin for children < 25 kg, 50 mg nitrofurantoin for children > 25 kg) (oral)

Control (16):
Matching placebo

Duration of intervention:
10 months

Washout period:
1 month

Outcomes

Rate of symptomatic UTI per catheterisation week (first period):
P:5/139; C:11/153

Rate of symptomatic UTI per catheterisation week (both periods):
P: 9/319; C: 19/309

At least 1 episode of definite symptomatic UTI (both periods):
P: 8/15; C: 11/15

NotesSymptomatic urinary tract infection defined as bacteriuria ≥ 10,000 cfu/ml with fever, abdominal pain, change in continence pattern, or change in colour or odour of urine
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot stated
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blind
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind
Incomplete outcome data (attrition bias)
All outcomes
Low riskOne child was discontinued from study because of a 4-week hospitalisation for torsion of an ovarian cyst
Selective reporting (reporting bias)Low riskAll outcomes related

Zegers 2011

  1. a

    cfu/ml = colony-forming units per millilitre

    C = control group

    P = prophylaxis group

    RCT = randomised controlled trial

    SMX = sulfamethoxazole

    T = temperature
    TMP = trimethoprim
    TMP-SMX = trimethoprim and sulfamethoxazole
    UTI = urinary tract infection

MethodsRCT parallel group
ParticipantsInclusion criteria: 176 children with known spina bifida performing clean intermittent catheterisation and using low dose chemoprophylaxis during preceding 6 months. They were allocated to continue or discontinue prophylaxis.
Interventions

Prophylaxis (88):

Trimethoprim, nitrofurantoin, cefuroxime, co-trimoxazole or a combination of prophylactics

Control (88)

No medicaments (placebo not used)

Duration:

18 months

Outcomes

At least one episode of febrile UTI

P:2/88; C: 4/88

Rate of asymptomatic significant bacteriuria per participants year

P: 3.64; C: 4.58

Rate of afebrile UTI per participants year

P: 1.07; C: 1.52

NotesSignifcant bacteriuria was defined as more than 103 cfu/ml. UTI was defined as significant bacteriuria, positive leukocyturia and clinical symptoms such as increasing incontinence and foul smell or cloudiness of urine with or without fever greater than 38.5 °C.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer- based random concealed allocation
Allocation concealment (selection bias)Low riskAdequate
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding
Blinding of outcome assessment (detection bias)
All outcomes
Low riskOutcome assessor blind
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo losses post randomisation. 38 (43%) switched back to prophylaxis group, analyses by intention- to- treat principles.
Selective reporting (reporting bias)Low riskAll outcomes related

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Chen 2005This study did not answer any of the questions posed in the review.
Clarke 2005The study had high risk of attrition bias.
Day 2003This study did not answer any of the questions posed in the review.
Jannelli 2007This study did not answer any of the questions posed in the review (short-term setting).
Kuhlemeier 1985The study methods were insufficiently described to judge eligibility.
Maynard 1984Authors did not present sufficient data to calculate incidence densities.
Mohler 1987bTrialists tried to answer three different questions. We did not include the following two comparisons because the study methods were too unclear: 'antibiotics if microbiologically indicated versus antibiotics when clinically indicated' and 'three days antibiotics versus 10 days antibiotics'.
Naik 2005This study did not answer any of the questions posed in the review (short-term setting).
Ratahi 2005Not published as an article.
Saint 2006This study did not answer any of the questions posed in the review (short-term setting).
Schumm 2008aNot published as an article.
Souto 2000This study did not answer any of the questions posed in the review.
Souto 2004This study did not answer any of the questions posed in the review.
Sublett 2007Not RCT.
Tang 2006This study did not answer any of the questions posed in the review (short-term setting).
Turi 2006Authors did not present sufficient data to calculate incidence densities.
Waites 2006This study did not answer any of the questions posed in the review.
Warren 1982bThis study did not address any of the questions posed in the review.