Urinary tract infections in spinal cord injury patients undergoing intermittent catheterization procedures

A prospective study was organized to assess whether feeding back infection rates to staff performing intermittent catheterization in spinal cord-injured patients would produce a fall in urinary tract infection rates. Over a 12 month period, infection rates for such procedures were reported to unit staff: reporting was combined with educational programs emphasizing aseptic catheterization techniques and the importance ofhandwashing. Overall infection rates for the 12 month period were 13.3 per 1000 days of intermittent catheterization - unchanged from the preceding six months (15.1 per 1000 days) by retrospective chart review. Likewise, there was no statistically significant downward trend during the prospective phase of the study. A wide variety of infecting organisms were found, of which Klebsiella species (39%), Escherichia coli (18%) and Enterobacter species (17%) were most common: most infections were asymptomatic. Patients with complete cord lesions at or above the sixth thoracic spinal cord segment (T6) had a much higher incidence of infection (73%) than those with incomplete lesions below T6 (33%).

I NTERM ITIENT CATHETERI ZATION PERFORMED UN DER sterile conditions is a well accepted method of managing acute spinal cord injury patients with a view to reducing the urinary infection rate.Nevertheless .bacteriuria is common and resultant upper tract infections remain an importa nt cause of morbidity in this population.Mos t a ttempts to reduce further infection rates h ave focused on the use of prophylactic urinary antiseptics or antibiotics (1 -4).The authors postulated that an important variable in the development of urinary tract infections (UTis) in this setting is the skill and attention of the individual performing the procedure, and that this could be influenced positively by reporting infection rates on a regula r basis to these operators.similar to programs reporting wound infection rates to practising surgeons (5).This report is a study of a program of regular reporting of infection rates to unit staff compared with a retrospectively obtained b aseline rate in the same population.

PATIENTS AND METHODS
The University of Alberta Hospita l is a 1355-b ed acu te tertiary care hospital providing a ll spinal cord rehabilitation services to northern Alberta.Once stabilized in an acute care setting, spinal cord -injured patients a r e transferred to the 26-bed rehabilitation unit in the same institution.The patient population on the rehabilitation unit is a mix of amputees and spinal cord-inj ured patients.the latter malting up 80% of the total population.The facility housing the rehabilitation unit was constructed in the 1950s, and system design has lin1ited the number of sinks that can be installed.A total of 10 sinks are present on the unit: three in the communal male washroom.three in the communal female washroom , two in the •utili ty rooms• (clean/ dirty).one in the hallway and one in the patient lounge.In January 1986 the staff of the spinal cord rehabilitation unit contacted the infection control unit with co ncerns that they were seeing excessive numbers of UTis.Consequ ently, the authors studied the problem by retrospective analysis ofUTis in the unit's pa tients over the preceding six months .and prospective analysis for the subsequent 12 months.Each patient studied h ad bladder dysfunction requiring more than one intermittent cath eterization each 24 h.Cath eterization was carried ou t by a trained group of orderlies (for m a le patients) or by registered nurses or registered nurse assistants (for femal e pa tients).all of whom were permanently assign ed to tl1e unit.Patients can perform their own catheterization as part of the training process and follow the same technique while in hospital .Sterile technique, including the use of a chlorhexidine solution for cleansing.was employed in the catheterization proced ure (6).
Urine c ult ures were submitted every two weeks for each patient and if a patient demonstrated clinical s ign s a nd /or symptoms suggestive of UTI.This wa rd practice remained unchanged tl1roughout tl1e retro spective and prospective phases of the study.UT! wai documented if: greater than 10 6 colonies/L were iso lated in a symptomatic patients; greater than 10< colonies/L were isolated on two separate occasions ir an asymptomatic patient; or greater tl1an 10< colonies/L were isolated on one occasion if eitl1e1 prophylactic or therapeutic antibiotics were prescribec without first repeating the urine culture (7) .
A c ulture was regarded as representing a relapse i the criteria for UT! were met, but the same bacteria species with identical antibiotic sensitivities was isolated fewer than 14 days after completion of a course ol antimicrobial therapy.Typing of isolates was not carried out to confirm strain identity.
Prophylactic antibiotics were not prescribed unde1 any circumstance.Asymptomatic UTis were treated when c ulture results became available witl1 oral agents based on sensitivity data.Patients with clinical signs a nd symptoms of systemic infection often had therapy initiated (usually ampicillin and gentan1icin) pending culture results, with subsequent modification of thera py.Asymptomatic infections were generally treated with five to seven days of therapy, and symptomatic infections 10 to 14 days , except in the case of relapses.which were usually treated with longer courses .
The retrospective study was carried out by review ol the medical records of all spinal cord-injured patients a dmitted to the same unit over a six month period (August 1985 to January 1986) .The prospective study was carried out over a 12 month period from February 1986 to January 1987.One of the authors (TK) visited the unit twice weekly to collect data by examining the medical records.Formal education sessions to review appropriate handwashing technique and aseptic catheterization procedure with health care workers were conducted on a weekly basis for the first montl1 during the prospective phase of the study.Continuing education occurred informally at the time of surveillance visits .Reports of infection rates were issued monthly to m edical and nursing staff, prominently displayed on the unit.and informally discussed witl1 various members of the health care team.

RESULTS
During the 18 montl1 study period, 69 patients were admitted for a total of 6400 days of intermittent catheterization procedures.Eighly-five infections were seen in 38 patients.Overall, 55% of patien ts experienced at least one UTI.When calculated as infections per 1000 days of intermittent catheterization procedures, the infection rate was 13.3: 12.4 in the prospective period and 15.1 in the retrospective period (P=0.44).Figure 1 gives the infection rate by montl1 for each period.Quarterly infection rates per 1000 days of intermittent catheterization procedures during the prospective period were 14.7 , 12.5, 11.0 and 11.1   1).Clinical signs and symptoms were infrequent (Table 2).Over one-half of infected patients were asymptomatic.and 75% h ad no clinical s igns of infection.Fever was the most common sign (22 .4%) .Urinalysis was abnormal in 51 of the 70 patients in whom it was obtained (50 with pyuria and eight with h e maturia).Fifty-six per cent of patients with abnormal and 26% with normal urinalysis had clinical signs or symptoms of infection.In only three cases was bacteremia documented (3.5%): blood cultures were not taken in every case.Infecting organisms were widely distributed among Gram -negative species: Klebsiella species (38.8%).Escherichia coli (18.4%),Enterobacter species (16 .5%).Proteus species (11.6%) and Pseudomonas aeruginosa (7 .8%).Serratia marcescens and Morganella morganii accounted for 0.9% each of infecting organisms.Grampositive organisms including Enterococcus jaecalis, coagulase negative staphylococci and Staphylococcus aureus accounted for 10.7% of infecting organisms.No clustering of cases was noted .There was no death related to infection.Despite the relative lack of severity of infections.heavy antibiotic use resulted.A total of 861 days of therapy were given either for acute infection (623 days) or relapse (238 days).One patient underwent prostatectomy for prostatitis and recurrent UT!.Infected patients spent an average of 148 days in hospital (range 50 to 405).whereas noninfected patients spent an average of25 days in hospital (range 20 to 93), probably also reflecting the severity of the underlying cord lesion.

DISCUSSION
The primary goal of this project, lo demonstrate a reduction of UT!s by reporting infection rates on a monthly basis to the ward staff involved in catheterization combined with ed u cational programs emphasizing aseptic catheterization techniques an d the importan ce of handwashing, was unsu ccessful.Since the a uthors clid not actually study the staffs cath eterization technique before and after the study was instituted, it is possible that there was no improvement in performance, resulting in no change in UT! rates.It is also possible that the study design contributed to this failure.The retrospective study m ay h ave underestimated the true infection rate.Since many infections a re asymptom atic, if urine specimens were less frequ ently subm itted for culture, fewer infection s would have been detected.The ward policy of submitting urine for c ul -ture remained unchanged throughout the study period: however , the authors• presence may have induced more frequ ent submission of specin1ens in asymptomatic patients.A more rigorous approach would have been to co ndu ct an entirely prospective study, but to witl1hold reporting the data for an initial period.It seems unlikely that extending the study would h ave changed th e res u lts , since there was no statistically significant downward trend in month -by-month infection rates .In any case, given the low short term morbiclity of these infections , a labour-intensive infection control program may not be justifiable on these grounds alone .
One implication of this failure to reduce infection rates m ay be that compliance with recommended catheterization technique was good and that witl1 this techniqu e there is an irreducible, minin1um number of infections which will occur.Certainly, the overall infection rate of 13 .3per 1000 patient days is very close to that seen by Rhane and Perkosh (1) (10.3 per 1000 patient d ays), in which prophylactic neomycin-polymyxin irrigant was also used, and lower than another study in which a rate of 19 per 1000 patient days with prophylactic antibiotics, and 65 per 1000 patient days witl1out prophylactic antibiotics, was found (4).Ra ther than being carried on the hands of the operators, organisms introduced into the bladder during endemic infections may derive from urethral and para-urethral flora (8).Brief clisinfection with chlorhexicline may be inadequate to eliminate these organisms prior to passing a catheter.
As in other studies (9) , tl1e a uthors found that most UTis in this setting are asymptomatic, and that little morbidity results.Only one p a tient required prostatic surge1y for recurrent UTI.Antibiotic use was h eavy.however, adcling significantly to hospitalization costs .Heavy antibiotic use may account for the greater number of antibiotic-resistant organisms isolated , particula rly Klebsiella and Enterobacter species, than reported from a large multi-institutional series of nosocomial UTis, which found E co li (32%) and enterococci (14%) to be predominant pathogens (10).
In summary, while having failed to d emonstrate the effectiveness of a program of regular reporting of infection rates for the purpose of reducing those rates , U1e a uthors h ave confirmed that intermittent catheterization is safe, and that most infection will b e seen in a s ubgroup of severely injured patients in whom recurrent infections , heavy antibiotic use and long hospi talization can be anticipated.

Figure 1 )
Figure 1) lrifection rates per 1000 patient days of intem1ittent catheterization procedures in 69 spinal cord -iryured patients at the University of Alberta Hospitals .Surveillance periods w ere August 1985 to January 1986 (retrospective) and February 1986 to January 1987 (prospective)

TABLE 1 Infection rate by site and co mple te ness of cord lesion Patients with at least Number of infections
lesions had a major impact on the overall infection rate , varying from 73% of patients with complete lesions at or a bove the sixth spinal cord segment (T6). to 33% of those below T6 with incomplete lesions (Table
CAN J INFECT D1s VOL 3 No 3 MAY/ JUNE 1992