A clinical evaluation of a sensor to detect blockage due to crystalline biofilm formation on indwelling urinary catheters

Authors


Abstract

Objective

To test the performance and acceptability of an early warning sensor to predict encrustation and blockage of long-term indwelling urinary catheters.

Patients and Methods

In all, 17 long-term indwelling catheter users, 15 ‘blockers’ and two ‘non-blockers’ (controls) were recruited; 11 participants were followed prospectively until catheter change, three withdrew early and three did not start.

Two sensors were placed in series between the catheter and the urine bag at catheter change. The sensor nearest the bag was changed at the same time as the bag change (weekly); the sensor nearest the catheter remained in situ for the duration of the catheter's life.

Bacteriology and pH determinations were performed on urine samples at each bag, sensor and catheter change. The colour of the sensors was recorded daily. On removal, each sensor and the catheter were examined for visible evidence of encrustation and blockage.

Participants were asked to keep a daily diary to record colour change and any other relevant observations and to complete a psychosocial impact of assistive devices tool at the end of the study. Participants and carers/healthcare professionals (when involved in urine bag or catheter change) were asked to complete a questionnaire about the sensor.

Results

Urease-producing bacteria were isolated from seven of the 14 patients (including early withdrawals; P. mirabilis in four, Morganella or Providencia in three).

In six of the seven patients the sensors turned blue-black; two of these were early withdrawals, two went to planned catheter change (one of these was recruited as a ‘non-blocker’) and three had catheter blockage.

The number of days of catheterisation before blockage was 22, 23 and 25 days, and the sensor changed colour within 24–48 h after insertion.

The urine mean (range) pH of the sensors that turned blue-black was 7.6 (5.5–9.0) and of the sensors that remained yellow 6.1 (5.1–7.5).

The sensor was generally well-received and was positive in the psychosocial assessment.

Conclusions

The sensor is a useful indicator of urine pH and of the conditions that lead to catheter blockage.

It may be particularly useful for new indwelling catheter users.

To be a universally acceptable predictor of catheter blockage, the time from sensor colour change to blockage needs to be reduced.

Introduction

Encrustation and blockage of indwelling (Foley) catheters has been reported to affect up to half of all long-term users [1]. This is a potentially serious problem that impacts on patient morbidity and health-related quality of life (HRQL) [2]. Blockage stems from infection with urease-producing bacteria, particularly Proteus mirabilis. Urease is a catalyst for the production of ammonia from urea causing an increase in urine pH [3]. Under these alkaline conditions, crystals of calcium and magnesium phosphate form, which, when combined with bacterial biofilm and debris, block the flow of urine down the catheter [4]. Patients have been subsequently categorised as ‘blockers’ or ‘non-blockers’ [1], with ‘blockers’ found to be significantly more often colonised with P. mirabilis and Providencia stuartii than ‘non-blockers’, and to excrete significantly more alkaline urine and lesser amounts of magnesium, urea and phosphate in their urine [4].

There is no reliable way of preventing or accurately predicting when blockage may occur [5]. Patients who are known ‘blockers’ may have frequent scheduled catheter changes in their care plan, but the unpredictability of blockage results in unnecessary or emergency changes, causing trauma to the patient and potentially avoidable healthcare costs [6].

An early warning sensor has been developed to predict catheter blockage that is caused by urease-producing bacteria. The sensor is designed to change colour from yellow to blue-black when the pH of the urine becomes alkaline, i.e. pH 6–8. Change in pH is used as a proxy indicator for imminent catheter blockage. An earlier prototype, a cellulose acetate and bromothymol blue composite strip held in place in the urine bag with a paper clip, was clinically assessed to establish proof of principle; this device demonstrated excellent sensitivity and gave prior warning of blockage on average 43 h in vitro [7] and 12 days in vivo [8].

The device has now been developed to make it suitable for large-scale manufacture. The modified sensor has a silicon-based composite strip housed in a small, clear polyvinyl chloride connector positioned between the catheter and drainage bag or valve, where it is easily visible but not intrusive. As urine drains into the bag it passes through the sensor. Results of in vitro testing confirmed function and gave an early warning signal 17–24 h before blockage [9]. The sensor is a Class 1 device manufactured by MBI (Wales) Ltd under license from Cardiff University.

The aim of the present pilot study was to evaluate the performance and acceptability of a modified sensor in clinical use. The aims were to gauge the longevity of the sensor, assess its sensitivity and specificity, observe the time interval between the colour change in the sensor and catheter blockage, and to establish user and healthcare professional (HCP)/carer perception of clinical relevance and usability.

Patients and Methods

The study was a pilot prospective observational study to record outcomes from 10 ‘blockers’ and two ‘non-blockers’ who would serve as controls. Ethical approval was obtained from the National Research Ethics Service (NRES) Committee North West Greater Manchester South (No. 10/H1003/121). Inclusion criteria were long-term suprapubic or urethral catheter use (≥6 months) and a plan to continue usage for a further 3 months. The ‘blockers’ were required to have a history of catheter blockage (patient records) during at least 3 of the last 6 months, and the ‘non-blockers’ to have a blockage free history for the previous 3 months prior to the study. Participants in receipt of antibiotic treatment ≤2 weeks of commencement of the study were excluded. Participants on prophylactic antibiotics were included. All participants were required to be able to give informed consent. Participants were requested to continue their normal regime and that no action should be taken in response to the sensor.

Each participant commenced the study on the date of a catheter change. To establish sensor longevity and thus the optimal time for changing the sensor, two sterile sensors were fitted in series between the catheter and the bag (Fig. 1). The sensor nearest the catheter (A) was left in situ for the duration of the catheter life; the sensor nearest the bag (B) was changed when the catheter bag was changed (usually weekly). In clinical practice only one sensor would be needed. The colour of the sensors (compared against a colour chart) was recorded daily by the participant or carer. A fresh urine sample was taken via the catheter bag and the pH of the urine recorded (pH meter, HANNA Instruments) by the research nurse each time sensor (B) was changed. Urine samples were transferred to universal boric acid containers for transport to the BioMed laboratory for bacterial culture. They were stored at <4 °C for ≤48 h before culture. This process was repeated until either the catheter blocked and was changed or a scheduled change took place. The catheters were also collected and sent to the laboratory for visual (photographed) assessment ofencrustation. Urine samples were subjected to a Combur test stick (Roche, Germany), to test for presence of nitrite in the urine. A serial dilution of the urine sample was then carried out using quarter strength Ringers solution (Oxoid, UK) as diluent. The dilution series was plated out onto Brilliance UTI agar and a selective medium for growth of Proteus [10]. After 18–24 h incubation, the bacteria were enumerated and, in the case of Brilliance UTI media, the colour of the colonies was noted, which allowed for a presumptive identification.

Figure 1.

Sensors A & B in situ.

At catheter change, both sensors were removed and participants and the HCPs/carers attending to the catheter changes were asked to complete a short questionnaire on the performance and acceptability of the sensor. The questionnaires required mostly ‘yes/no’ answers, but participants were invited to add free text comments at all stages. Comments pertaining to the sensor made verbally by participants, carers or HCPs during the research nurse's visits were recorded on the Clinical Report Forms. Participants were also asked to complete the Psychosocial Impact of Assistive Devices Scale (PIADS), a 26-item self-rating scale designed to provide a reliable and valid measure of how users perceive the impact of assistive devices on their HRQL and sense of well-being [11]. The scale measures from –3 (negative impact) to +3 (positive impact).

Results

In total 17 participants were consented between October 2012 and March 2013 from Southmead Hospital, Bristol and from the local community via district nurses and through direct communication. The two ‘non-blockers’ (ES01, ES05), both male and with suprapubic catheters, consented and completed the study. Fifteen ‘blockers’ consented; three (ES04, ES06, ES12) did not start; three (ES10, ES11, ES17) withdrew early; nine (ES02, ES03, ES07, ES08, ES09, ES13, ES14, ES15, ES16) completed the study (i.e. to catheter change). Of the 12 ‘blockers’ who started, five were female and seven were male; five had urethral and seven suprapubic catheters; participant ES08 took prescribed prophylactic antibiotics and ES15 used a nitrofurazone-release catheter (Rochester Medical); nine were on a regime of regular catheter changes of between 2 and 12 weeks (mean 5.4 weeks). Participant ES07 was receiving regular weekly bladder wash-outs with SubyG solution. Three-quarters of respondents had significant mobility issues; eight used a wheelchair and two required a hoist to get out of bed.

Sensor Longevity and Performance

The two ‘non-blockers’ had scheduled catheter changes at 54 and 70 days, respectively. There was no evidence of catheter blockage during this period; no presence of P. mirabilis and both sensors (A & B) remained yellow throughout. Participant ES01 was found to be positive for the urease-producing organisms Morganella or Providencia spp. A mean (range) urine voiding pH (pHv) of 6.4 (6.1–6.8) and 5.5 (5.1–6.0) for participants ES01 and ES05 respectively, were recorded. Both participants had high levels of bacteriuria (105 colony-forming units[cfu]/mL).

Of the ‘blockers’ only four (ES02, ES08, ES11, ES17) were positive for P. mirabilis and only one of these (ES02) went to catheter blockage; ES08 had a planned catheter change after 50 days, and ES11 and ES17 both withdrew from the study early. In all four, the sensors changed from yellow to blue-black. Participants ES13 and ES15 also went to catheter blockage and the sensors changed from yellow to blue-black; in both patients either Morganella or Providencia spp. were present in the urine. One participant (ES03) had his catheter changed due to ‘by-passing’ (leakage around the outside of the catheter) but no urease-producing bacteria were present and the sensor did not change colour. However, his pHv was relatively high, at a mean (range) of 7.3 (6.3–8.1). Participants ES07, ES14 and ES16 went to planned catheter change and ES09 underwent catheter change during a trial-without-catheter (TWOC). No urease-producing bacteria were recorded as being present. The sensors remained yellow throughout for participants ES07, ES09 and ES16, but changed blue-black for participant ES14. Participant ES10 withdrew early and there was no evidence of urease-producing bacteria or sensor colour change. The results are summarised in Table 1.

Table 1. Summary of performance results
Patient IDHistorySexCatheter routeCatheter typeSize, FDays of catheterisationReason for catheter changeBacteriaUrease-producing bacteriaSensor colourMax pHMean pH
  1. NF, nitrofurazone.
ES01Non-blockerMaleSuprapubicSilicone1654PlannedMorganella or Providencia, PseudomonasYesYellow6.86.4
ES05Non-blockerMaleSuprapubicSilicone1670PlannedEnterococcus, PseudomonasNoYellow6.05.5
ES02BlockerFemaleUrethralSilicone1225BlockedP. mirabilisYesBlue-black8.28.1
ES03BlockerMaleUrethralSilicone1642By-passingE. coli, Pseudomonas, StaphylococcusNoYellow7.57.2
ES07BlockerFemaleUrethralSilicone (wash outs)1449PlannedE coli, PseudomonasNoYellow5.55.4
ES08BlockerMaleSuprapubicSilicone (prophylactic antibiotics)1650PlannedP. mirabilis, E. coli, PseudomonasYesBlue-black8.67.4
ES09BlockerMaleUrethralHydrogel-coated latex1213TWOCE. coli, PseudomonasNoYellow5.35.2
ES10BlockerFemaleUrethralSilicone147Withdrew earlyE. coli, PseudomonasNoYellow5.85.8
ES11BlockerMaleSuprapubicSilicone168Withdrew earlyP. mirabilisYesBlue-black9.07.3
ES13BlockerMaleSuprapubicHydrogel-coated latex1622BlockedE. coli, Pseudomonas, Staphylococcus, Morganella or ProvidenciaYesBlue-black7.77.7
ES14BlockerFemaleSuprapubicSilicone1442PlannedPseudomonas, Staphylococcus, Enterococci?NoBlue-black7.77.4
ES15BlockerMaleSuprapubicNF-release1823BlockedPseudomonas, Morganella or Providensia, StaphylococcusYesBlue-black8.07.5
ES16BlockerMaleSuprapubicSilicone1653PlannedStaphylococcus, PseudomonasNoYellow7.06.5
ES17BlockerFemaleSuprapubicSilicone187Withdrew earlyP. mirabilis, Staphylococcus, PseudomonasYesBlue-black8.48.1

For analysis, the initial baseline pH readings were excluded as these did not relate to the colour of the sensor. All other readings were included. The sensor changed to blue-black with pHv readings from 5.5 to 9.0 (mean 7.6) and remained yellow with pHv readings from 5.1 to 7.5 (mean 6.1) (Fig. 2). Apart from a single recording of the sensor turning blue-black at pHv 5.5, the range in which the sensor either remained yellow or turned blue-black was between pHv 6.5 and 7.5. With one false-negative (ES01) and one false-positive (ES14), the specificity and sensitivity of the sensor to the presence of urease-producing organisms were both 0.86. Times to catheter blockage occurred 25, 22 and 23 days after insertion for participants, ES02, ES13 and ES15, respectively. Sensors A and B turned blue-black between 3 and 5 days after initial insertion. Sensor A stayed blue-black throughout and sensor B changed from yellow to blue-black within 24–48 h after each sensor change. Participants ES07, ES08, ES14 and ES16 had planned catheter changes between 42 and 53 days. In participants ES08 and ES14, sensor A turned blue-black after 2–5 days and sensor B changed within 24–48 h after each sensor change. In participants ES07 and ES16 both sensors remained yellow throughout. The data indicate a distinction in the number of days of catheter life between catheters that blocked and those that did not block (Fig. 3).

Figure 2.

Sensor colour and pHv.

Figure 3.

Days in the study and reason for catheter change.

All participants had high levels of bacteriuria (105 cfu/mL). In all, 12 participants had positive cultures for the Gram-negative bacteria Pseudomonas spp., four for P. mirabilis, six for E. coli and three for Morganella or Providencia spp. (the culture method did not allow for distinction between them); seven participants had positive cultures of the Gram-positive bacteria Staphylococcus spp. and two for Enterococcus spp.

Acceptability

In total 12 participants completed the patient sensor evaluation questionnaire (Table 2). Most respondents found the sensor to be easily accessible and acceptable in appearance; the colour change was easy to monitor and the sensor was thought to be a good idea.

Table 2. Participant evaluation summary
 n/N
YesNoNot sureNo response
Was the sensor easily accessible?11/12001/12
Was the colour of the sensor easy to read?10/122/1200
Did the sensor interfere with your normal daily life?7/125/1200
Was the appearance of the sensor acceptable?9/123/1200
Did the sensor leak at the connection point to the catheter or drainage bag?5/127/1200
Did the sensor make you feel less anxious about blockage?3/127/122/12 
Do you think that having a sensor to tell you that a blockage is about to occur is a good idea?9/1203/120
When available, will you choose to use the sensor?7/122/123/120

Those respondents who reported that the sensor had interfered with their daily lives found the additional length of the two sensors fitted in series (for research purposes only) caused complications when dressing and made it difficult for them to keep the catheter hidden. This had contributed to at least one early withdrawal from the study. Most respondents thought that a single sensor (normal use) would be more discreet. Nearly half of respondents had some leakage aroundthe sensor, but this was infrequent and usually resolved quickly.

Before evaluating the sensor, over half of participants reported being anxious about catheter blockage; after the trial period, one-quarter of the respondents thought having the sensor would make them less anxious. Participants expressed concern that the time between the sensor changing colour and catheter change/blockage could be several weeks and that the sensor ‘won't make it block less’. Positive comments included: ‘[the sensor] would [make me] feel more confident’ and ‘[it] would be nice to know [if the catheter is going to block]’.

In all, 11 participants completed the PIADS. The sensor was found to have a small positive impact on user psychosocial well-being with scores of +0.43 for ‘Competence’ (reflecting perceived functional capability, independence and performance), +0.41 for ‘Adaptability’ (reflecting inclination or motivation to participate socially and take risks) and +0.27 for ‘Self-esteem’ (reflecting self-confidence, self-esteem, and emotional well-being).

Six HCPs/carers completed the sensor evaluation questionnaire (Table 3). Three reported that the sensors had leaked; on further questioning by the research nurse this appeared to be largely due to problems with fitting the two sensors together and was a temporary issue. All thought the appearance was acceptable and would be discreet if one sensor only was in place. Opinion was divided as to whether the sensor would be a good indicator of catheter blockage, although most would recommend it if the time between the sensor changing colour and blockage was shorter. The colour change was deemed easy to identify by most carers and users, and those that commented said that the instructions for use were good.

Table 3. HCP/carer evaluation summary
 n/N
YesNoNot sureNo response
Was the sensor easy to change?4/601/61/6
Was it easy to identify colour change?4/601/61/6
Did the sensor leak?3/62/61/6 
Do you think the appearance of the sensor is acceptable?6/60  
Do you think the sensor is discreet?5/6 1/6 
Were the instructions for use clear for use?3/601/62/6
Do you think the sensor is a good indicator of imminent blockage?3/603/6 
Would you recommend the sensor to long term indwelling catheter users?4/602/6 

Discussion

The number of participants in this pilot study is small and no statistical analysis has been performed. However, the data offers some strong indications as to the performance and acceptability of the sensor.

As expected, all participants had high levels of bacteriuria. However, despite having a history of catheter blockage, only three of the 12 ‘blockers’ had catheter blockage during the study period, and only four of the 12 were positive for P. mirabilis, two of whom withdrew early, one (ES17) having catheter blockage shortly after withdrawal. These results contrast with the clinical evaluation of the earlier sensor [8] in which 15 of 20 participants were positive for P. mirabilis (75%), and 12 had catheter blockage (60%). The reason for the difference in colonisation and subsequent blockage is difficult to ascertain without detailed history and could simply be due the low number of participants in the present study. Both studies confirm the high degree of variability in time to blockage of the catheter, reported as taking from 2 to 98 days [12]. In the present study, the time between sensor colour change and catheter blockage was >18 days. This is on average 6 days longer than in the clinical evaluation of the first prototype but within the range of 2 to 34 days recorded [8]. This difference may be attributed to the few participants who had blockage during the present study, or to differences in the sensor device construction and/or its location; the initial prototype was fastened to the inside of the catheter bag, which was found to give a stronger signal than positioning at the catheter-drainage bag junction [7].

The longevity of the sensor was difficult to ascertain, as the sensors indicated a rise in pH, usually due to the presence of urease-producing bacteria, but blockage did not occur until several weeks later. The sensor (A), which was in place for the duration of the catheter life, usually changed in response to raised pHv after a few days and remained blue-black regardless of fluctuations in pH. Sensors (B) that were replaced weekly usually changed blue-black at ≤24 h in response to a raised pHv.

The mean time between sensor change and blockage was considerably longer in human trials than in the in vitro model; 43 h [7] vs 12 days [8] for the first sensor prototype and 17–24 h [9] vs ≥19 days for the modified sensor. The model used in both in vitro studies is intentionally designed to accelerate the blockage process and cannot adequately simulate all the factors that impact on urine composition. Thus, although the models have been used to develop and test the sensors, they cannot accurately predict the time between sensor colour change and blockage in vivo.

The modified sensor appears to be an accurate predictor of raised pHv and of the presence of urease-producing bacteria, but the time from sensor colour-change to catheter blockage is arguably too long to be clinically useful. Whether the sensor is measuring pHv or the pH of the biofilm is not certain, but the time between sensor indication and blockage would suggest that the relationship between pH and the encrustation process may be too complex or capricious for pHv to be a single proxy measure for impending blockage. Several studies have investigated the relationship between pHv, the presence of urease-producing micro-organisms and propensity for catheters to encrust [13-15]. Hedelin et al. [14] found the pH that divided ‘blockers’ from ‘non-blockers’ to be ≈6.8, but subsequent studies have shown that catheters do block at a pH <6.8 and do not block at a higher pH. More recent studies have looked at other factors that affect pH and encrustation, particularly that of nucleation pH (pHn), i.e. the pH at which crystals of struvite and apatite precipitate out of solution. It has been reported in vitro [16] and in a clinical study [17], that the nearer the pHn is to the pHv the greater the propensity for stone formation and catheter encrustation. The study by Suller et al. [18] (2005) describes the effect of fluid intake and citrate as dilution and chelating agents respectively, which raise the pHn thus also raising the pHv at which crystal deposition occurs. This strategy for pH modulation has been shown to be effective in vivo [19] as well as in vitro [20]. The relationship between pH and encrustation is further complicated by the variation in ability to produce urease within the Proteus genus; P. mirabilis and P. vulgaris species are the most dominant, but there is also variability with sub-species and type [21], a factor which could contribute to the premise of ‘fast’ or ‘slow’ encrusters [12]. These factors preclude a simple relationship between pHv and catheter blockage, and regular urine pH sampling as part of catheter care has been discredited [22, 23].

Most patients and HCPs/carers thought the sensor could help alleviate the anxiety associated with catheter blockage and the positive PIADS is encouraging, especially when compared with a negative PIADS for Foley catheter users [24]. Although some participants and HCPs/carers were concerned about the additional length imposed by the two sensors during the study, in practice this is unlikely to be a problem, as only one sensor would be in place. However, due to the length of time between sensor colour change and catheter blockage the sensor as an indicator of imminent catheter blockage was questioned.

If the sensor can be made to respond closer to the time of catheter blockage, the sensor could potentially be a useful clinical indicator. Whether this can be achieved through measurement of pH alone may need to be considered.

Acknowledgements

The clinical study was funded by a charity supporting medical research. The authors are grateful to Dr Paul White, University of West of England for his advice on data analysis and to Dr Nicola Morris, North Bristol NHS Trust, for additional microbiology advice. Our thanks go to Bristol Community Health, South Gloucester Community team and North Somerset Community Partnership for their invaluable help with patient recruitment, and of course to the catheter users, carers and HCPs who participated. We are also grateful to the University of Cardiff and MBI Ltd for the supply of sensors and to Professor Roger Feneley for his continued support in the development of this product.

Conflict of Interest

None declared.

The original pH sensor was developed in collaboration with the BioMed Centre, Bristol Urological Institute, North Bristol NHS Trust. The patent is owned by Cardiff University and licensed to MBI (Wales) Ltd. A royalty agreement is in place with North Bristol NHS Trust if the device is fully commercialised.

Abbreviations
cfu

colony-forming units

HCP

healthcare professional

HRQL

health-related quality of life

pHn

nucleation pH

pHv

urine voiding pH

PIADS

Psychosocial Impact of Assistive Devices Scale

TWOC

trial-without-catheter

Ancillary