EfifiSh J011rr~rlOf U ~ O U(1979), 51, 317-320
Faecal and lntroital Bacteria and Urinary Tract Infection B. RICHARDS and E. MARY COOKE Departments of Urology and Microbiology, York District Hospital, York
Summary-The faecal, introital, and urinary bacterial flora have been studied in 3 normal women and 5 women with recurrent urinary infection. In the normal women urinary abnormalities were uncommon, but the patients regularly had episodes of bacteriuria and pyuria, only a quarter of which were symptomatic. Symptoms tended to be associated with high white cell excretion rates and with the longer episodes. lntroital colonisation was heavier and more frequent in the patients than in the control subjects. Organisms recovered from the urine had previously colonised the introitus in most cases. It appears that symptomatic episodes constitute only a small part of the disease process in patients with recurrent urinary tract infection.
Urinary tract infection in women is a distressingly common condition. Most of the infections are ascending (Hinman, 1966) and the organisms originate in the patient’s own bowel. It has been suggested that one of the factors that predisposes to the development of infection is colonisation of the vaginal vestibule with urinary pathogens (Stamey et al., 1971; Stamey, 1973), though other workers have been unable to demonstrate a close relationship between introital enterobacteria and urinary infection (O’Grady et al., 1970; Marsh et al., 1972). This study was undertaken in an attempt to elucidate the relationship between the faecal flora, introital colonisation and the development of urinary tract infection.
Patients and Methods Five women, aged 21 to 43 years, were studied; each had suffered at least 2 attacks of bacterial cystitis in the 3 months before the start of the investigation. All had normal excretion urograms and in all the residual urine was less than 2 ml (Barnard et al., 1973). The 3 controls, aged between 21 and 39 years, had no history of urinary tract infection. All of the individuals taking part in the investigation gave twice-weekly stool samples, twiceReceived 10 October 1978. Accepted for publication 21 February 1979.
317
daily introital swabs and daily early morning urine samples. The investigation lasted on average 89 days for each patient (range 67 to 107 days). Specimens of faeces were inoculated onto MacConkey agar and examined for coliform organisms, enterococci and micrococci. Five colonies of each colonial type of coliform were identified by their biochemical reactions (Cooke et al., 1969) and examined for haemolysin production (Cooke, 1968). Selected strains of E. coli, having biochemical and haemolytic characteristics similar to strains subsequently found in the urine, were serotyped using 150 0 antisera and 50 H antisera (Bettelheim and Taylor, 1969). Catalasenegative Gram-positive cocci which were isolated on MacConkey agar were considered to be enterococci. Micrococci were identified by their biochemical reactions (Baird-Parker, 1966). Swabs were taken from the vaginal introitus morning and evening using swabs moistened in nutrient broth. These were wiped around the external meatus, with the labia held apart, and were immediately placed in Stuart’s transport medium. They were inoculated onto MacConkey and blood agar and the organisms were examined in the same way as those from the faeces. Where several colonial types of E. coli were isolated, serotyping was carried out only on those with biochemical and haemolytic reactions similar to the organisms subsequently recovered from the urine. The number of organisms recovered from
318
BRITISH JOURNAL OF UROLOGY
the introital swab was expressed semi-quantitatively (O’Grady et al., 1970). Clean catch early morning urine specimens were obtained daily and refrigerated until examined. Cells were counted without centrifugation. The urine was inoculated onto MacConkey agar using a standard loop. The organisms isolated were examined in the same way as those from faeces except that only one colony of each colonial type was examined. Results Urinary Tract Infection The urine was considered to be abnormal if more than 10 pus cells/mm3 were present or if there was a significant bacteriuria (> 10S/ml). Episodes of pyuria or bacteriuria were relatively common in the patients, occurring on average once every 10 days. There was little difference between the patients in the frequency of episodes, the range being from an average of one episode in 7 days to one in 12 days. Of the 44 episodes occurring during 446 patient days, 38 were pyuric and 30 bacteriuric. In only 13 episodes did patients suffer from symptoms of infection. These symptomatic episodes were distributed between the 5 patients as follows: 5, 3, 2, 2, 1. The 30 episodes of bacteriuria were associated with the following organisms: E. coli 16, Strep. faecalis 8 and micrococci 6. In the controls there were 2 episodes of pyuria occurring in 2 individuals. There were no episodes of bacteriuria and no symptoms of infection (Table 1). In the patients, symptoms tended to be associated with a high white cell excretion rate and with longer episodes of pyuria or bacteriuria (Table 2). Seven of the bacteriuric episodes consisted of only one infected specimen and none of these was associated with symptoms. Table 1 The Incidence of Pyuria, Bacteriuria and Symptoms of Urinary Tract Infection in Patients and Controls
No. of episodes of pyuria
No. of persons
No. of palienl days
3
191
2
0
0
Patients with UTI 5
446
38
30
13
Controls
No. of episodes of bcrrreriuria
No. of episodes asmialed with symploms of
UTI
Table 2 The Association Between the Duration of Bacteriuria, the Duration and Severity of Pyuria and the Development of Symptoms of Urinary Tract Infection
Pyuria of 1-3 days Pyuria of 4 days or more Bacteriuria of 1-3 days Bacteriuria of 4 days or more Pyuria of 10-60cells/mm3 Pyuria of more than 60 cells/mm3
No. of
No. of
episodes associated with symptoms of
episodes not associated with symptoms of
UTI
UTI
5
4 5
23 2 21 0
5 8
19 6
8
Introital Colonisation The colonisation of the introitus with E. coli was heavier and more frequent in the patients than in the controls. E. coli was recovered from 16% of introital swabs from patients and from 6% of swabs from the controls. A heavy growth was obtained from 7% of the patients’ swabs and 1 Yo of the swabs from the controls. Urinary and Introital Bacteria The results of the examination of introital swabs in the days preceding a bacteriuric episode were available in 28 of 30 episodes. Organisms considered to be similar when identified biochemically were present in the introitus and the urine in 23 of these 28 episodes. On 16 occasions E. coli was found in the introitus before appearing in the urine. On 12 occasions E. coli from both sources were serotyped and in only one case was the serotype different. The period of introital colonisation preceding a bacteriuric episode was variable. Eight of the episodes of bacteriuria followed a period of introital colonisation exceeding 5 days’ duration. Urinary and Faecal Bacteria For 28 bacteriuric episodes faecal specimens were obtained during the 6 days preceding the episode and for 11 of the episodes 2 specimens were obtained during this time. Similar organisms, when identified as described, were found in 19 of the 28 episodes. The urinary organism was E. coli in 16 of the episodes; serotyping was carried out in 11 of these and the similarity of the organisms
FAECAL AND INTROITAL BACTERIA AND URINARY TRACT INFECTION
was confirmed in 5. Micrococci were isolated from the urine on 6 occasions; they were never isolated from the faeces.
Haemolysin Production Eight per cent of isolates of E. coli from the faeces were haemolytic. There was no difference between the controls and the patients in the faecal carriage rate of haemolytic E. coli. Eighteen per cent of isolates from the introitus were haemolytic. This comprised 21% of the 1166 isolates from patients and 2% of the 150 isolates from the controls. No E. coli were isolated from the urine of the controls. Thirty-four per cent of the E. coli isolates from the urine of the patients were haemolytic.
319
obtained that introital colonisation may be of importance in the development of urinary tract infection. The results presented here support this view as introital colonisation was more frequent and heavier in the women with recurrent urinary infection that in the controls. In addition, the previous isolation of the infecting organism from the introitus was regularly demonstrated, as was the development of infection following a period of introital colonisation of more than 5 days’ duration. Although it is generally accepted that the bowel is the source of many of the organisms causing urinary infection, we were unable t o demonstrate in the faeces the serotype of E. coli which subsequently caused infection in 6 of the 11 patients in whom this was attempted. The reasons for this are not clear. Many of the investigations demonstrating faecal carriage of the infecting strain have been carried out subsequent t o the development of the infection (Gruneberg et a/., 1968; Bettelheim et al., 1971) and few prospective studies have been done. It has been shown that under certain conditions the faecal coliform population may change rapidly (Cooke et al., 1969) and although specimens of faeces were obtained twice-weekly it may be that the infecting strain was present in the faeces for such a short time before the development of the infection that it was not detected. The results for haemolysin production are similar to those obtained previously (Cooke and Ewins, 1975). The introital isolates from the patients were commonly haemolytic (2 19’0)whereas those from the controls were not (2070). Thirtyfour per cent of the urinary infecting strains were haemolytic. It has been shown (Cooke and Ewins, 1975) that introital strains that subsequently cause infection are more frequently haemolytic than those which d o not and the present findings support this view. This study of a small number of patients provides evidence that women with bacteriuria cannot be divided clearly into symptomatic and asymptomatic groups, as patients with symptoms also undergo asymptomatic episodes of bacteriuria and pyuria which outnumber the symptomatic episodes.
Discussion The initial aim of the investigation was to study in detail the relationship between introital colonisation and the development of urinary infection, but the finding which emerged most clearly was the frequency of asymptomatic infection in patients who were also suffering from recurrent symptomatic disease. Asymptomatic bacteriuria in non-pregnant women has been studied and it has been shown (Asscher et al., 1972) that symptomatic infection develops more frequently among bacteriuric than control subjects. In this investigation we have shown that in women with recurrent symptomatic infection the symptomatic episodes constitute less than a third of the episodes of bacteriuria and pyuria. Although symptoms were most frequently associated with bacteriuria and pyuria of more than 3 days’ duration and with pyuria of more than 60 cells/mm3, the possibility arises that these very frequent attacks of infection, which are untreated because they are asymptomatic, may be of importance in the ldevelopment of the disease. Radiological disease appears t o be more common in wholly asymptomatic patients (O’Grady el al., 1972). From the small number of patients studied here it appears that the distinction between asymptomatic and symptomatic patients may not be a clear one, as symptomatic patients also have many asymptomatic episodes. The relationship between introital colonisation and the development of urinary tract infection has Acknowledgements been studied by a number of workers (O’Grady We thank Mr L. Wilson for expert technical assistance, and et al., 1970; Stamey et al., 1971; Marsh et al., Dr K. Bettelheim and Mrs M . Chandler for typing the strains 1972; Stamey, 1973) and some evidence has been of E. coli.
320
BRITISH JOURNAL OF UROLOGY
Hinman, F. (1966).Mechanisms for the entry of bacteria and the establishment of urinary infection in female children. Asscher, A. W., Chick, S., Redford, N., Waters, W. E., Journal of Urology, 96, 546-550. Sussman, M., Evans, J. S., McLachlan, M. S. F. and Marsh, F. P., Murray, M. and Panchamia, P. (1972).The Williams, J. E. (1972). Natural history of asymptomatic relationship between bacterial cultures of the vaginal bacteriuria (ASB) in non-pregnant women. In Urinary introitus and urinary infection. British Journal of Urology, Tract Infection, ed. Brumfitt, W. and Asscher, A. W. 44,368-375. pp. 51-61.London: Oxford University Press. O’Crady, F. W., Charlton, C. A. C., Kelsey Fry, I., Balrd-Parker, A. G. (1966). Methods for classifying StaphylMcSherry, A. and Cattell, H’. R. (1972).Natural history of cocci and Micrococci. In Idenrification Methods for Microintractable ‘cystitis’ in women referred to a special clinic. biologists, ed. Gibbs, B. M. and Skinner, F. A. Pp. 59-64. In Urinary Tract Infedon, ed. Brumfitt, W.and Asscher, London: Academy Press. A. W . Pp. 81-91.London: Oxford University Press. Barnard, D., Bastable. J. R. G . and Richards, B. (1973). O’Grady, F. W., McSherry, M. A., Richards, B., O’Farreil, Estimation of residual urine volume with 1311-Hippuran. S. M. and Cattell, W. R. (1970). lntroital enterobacteria, British Journal of Urology, 45, 408-414. urinary infection and the urethral syndrome. Lancet, 2, Bettelheim, K. A., Dulake, C. and Taylor, J. (1971). Post1208-I210. operative urinary infections caused by Escherichia coli. Stamey, T. A. (1973). The role of introital enterobacteria Journal of Clinical Pathology, 24,442-443. in recurrent urinary infections. Journal of Urology, 109, Bettelheim, K. A. and Taylor, J. (1969). A study of Escher467-472. ichia coli isolated from chronic urinary infection. Journal Stamey, T. A., Timothy, M., Mlllar, M. and Mihara, G. of Medical Microbiology. 2, 225-236. (1971). Recurrent urinary infections in adult women: the Cooke, E. M. (1968). Properties of strains of Escherichia role of introital enterobacteria. California Medicine, 115, coli isolated from the faeces of patients with ulcerative 1-19. colitis, patients with acute diarrhoea and normal persons. Journal of Pathology and Bacteriology, 95, 101-113. Cooke, E. M. and Ewins, S. P. (1975).Properties of strains of Escherichia coli isolated from a variety of sources. The Authors Journal of Medical Microbiology, 0, 107-1I I. Cooke, E. M., Ewins, P. S. and Shooter, R. A. (1969). B. Richards, MD, FRCS, Consultant Urologist. Changing faecal population of Escherichia coli in hospital E. Mary Cooke, BSc, MD, MRCPath, Consultant Microbiologist, York. Now Professor of Clinical Microbiology, medical patients. British Medical Journal, 4,593-595. Institute of Pathology, Leeds. Gruneberg, R. N., Leigh, D. A. and Brumfitt, W. (1968). Escherichia coli serotypes in urinary tract infection: studies in domiciliary, ante-natal and hospital practice. In Urinary Tract Injection, ed. O’Grady, F. and Brumfitt, W. Requests for reprints to: B. Richards, Department of Urology, York District Hospital, Wigginton Road, York YO3 7HE. Pp. 68-69.London: Oxford University Press.
References
Faecal and lntroital Bacteria and Urinary Tract Infection B. RICHARDS and E. MARY COOKE Departments of Urology and Microbiology, York District Hospital, York
Summary-The faecal, introital, and urinary bacterial flora have been studied in 3 normal women and 5 women with recurrent urinary infection. In the normal women urinary abnormalities were uncommon, but the patients regularly had episodes of bacteriuria and pyuria, only a quarter of which were symptomatic. Symptoms tended to be associated with high white cell excretion rates and with the longer episodes. lntroital colonisation was heavier and more frequent in the patients than in the control subjects. Organisms recovered from the urine had previously colonised the introitus in most cases. It appears that symptomatic episodes constitute only a small part of the disease process in patients with recurrent urinary tract infection.
Urinary tract infection in women is a distressingly common condition. Most of the infections are ascending (Hinman, 1966) and the organisms originate in the patient’s own bowel. It has been suggested that one of the factors that predisposes to the development of infection is colonisation of the vaginal vestibule with urinary pathogens (Stamey et al., 1971; Stamey, 1973), though other workers have been unable to demonstrate a close relationship between introital enterobacteria and urinary infection (O’Grady et al., 1970; Marsh et al., 1972). This study was undertaken in an attempt to elucidate the relationship between the faecal flora, introital colonisation and the development of urinary tract infection.
Patients and Methods Five women, aged 21 to 43 years, were studied; each had suffered at least 2 attacks of bacterial cystitis in the 3 months before the start of the investigation. All had normal excretion urograms and in all the residual urine was less than 2 ml (Barnard et al., 1973). The 3 controls, aged between 21 and 39 years, had no history of urinary tract infection. All of the individuals taking part in the investigation gave twice-weekly stool samples, twiceReceived 10 October 1978. Accepted for publication 21 February 1979.
317
daily introital swabs and daily early morning urine samples. The investigation lasted on average 89 days for each patient (range 67 to 107 days). Specimens of faeces were inoculated onto MacConkey agar and examined for coliform organisms, enterococci and micrococci. Five colonies of each colonial type of coliform were identified by their biochemical reactions (Cooke et al., 1969) and examined for haemolysin production (Cooke, 1968). Selected strains of E. coli, having biochemical and haemolytic characteristics similar to strains subsequently found in the urine, were serotyped using 150 0 antisera and 50 H antisera (Bettelheim and Taylor, 1969). Catalasenegative Gram-positive cocci which were isolated on MacConkey agar were considered to be enterococci. Micrococci were identified by their biochemical reactions (Baird-Parker, 1966). Swabs were taken from the vaginal introitus morning and evening using swabs moistened in nutrient broth. These were wiped around the external meatus, with the labia held apart, and were immediately placed in Stuart’s transport medium. They were inoculated onto MacConkey and blood agar and the organisms were examined in the same way as those from the faeces. Where several colonial types of E. coli were isolated, serotyping was carried out only on those with biochemical and haemolytic reactions similar to the organisms subsequently recovered from the urine. The number of organisms recovered from
318
BRITISH JOURNAL OF UROLOGY
the introital swab was expressed semi-quantitatively (O’Grady et al., 1970). Clean catch early morning urine specimens were obtained daily and refrigerated until examined. Cells were counted without centrifugation. The urine was inoculated onto MacConkey agar using a standard loop. The organisms isolated were examined in the same way as those from faeces except that only one colony of each colonial type was examined. Results Urinary Tract Infection The urine was considered to be abnormal if more than 10 pus cells/mm3 were present or if there was a significant bacteriuria (> 10S/ml). Episodes of pyuria or bacteriuria were relatively common in the patients, occurring on average once every 10 days. There was little difference between the patients in the frequency of episodes, the range being from an average of one episode in 7 days to one in 12 days. Of the 44 episodes occurring during 446 patient days, 38 were pyuric and 30 bacteriuric. In only 13 episodes did patients suffer from symptoms of infection. These symptomatic episodes were distributed between the 5 patients as follows: 5, 3, 2, 2, 1. The 30 episodes of bacteriuria were associated with the following organisms: E. coli 16, Strep. faecalis 8 and micrococci 6. In the controls there were 2 episodes of pyuria occurring in 2 individuals. There were no episodes of bacteriuria and no symptoms of infection (Table 1). In the patients, symptoms tended to be associated with a high white cell excretion rate and with longer episodes of pyuria or bacteriuria (Table 2). Seven of the bacteriuric episodes consisted of only one infected specimen and none of these was associated with symptoms. Table 1 The Incidence of Pyuria, Bacteriuria and Symptoms of Urinary Tract Infection in Patients and Controls
No. of episodes of pyuria
No. of persons
No. of palienl days
3
191
2
0
0
Patients with UTI 5
446
38
30
13
Controls
No. of episodes of bcrrreriuria
No. of episodes asmialed with symploms of
UTI
Table 2 The Association Between the Duration of Bacteriuria, the Duration and Severity of Pyuria and the Development of Symptoms of Urinary Tract Infection
Pyuria of 1-3 days Pyuria of 4 days or more Bacteriuria of 1-3 days Bacteriuria of 4 days or more Pyuria of 10-60cells/mm3 Pyuria of more than 60 cells/mm3
No. of
No. of
episodes associated with symptoms of
episodes not associated with symptoms of
UTI
UTI
5
4 5
23 2 21 0
5 8
19 6
8
Introital Colonisation The colonisation of the introitus with E. coli was heavier and more frequent in the patients than in the controls. E. coli was recovered from 16% of introital swabs from patients and from 6% of swabs from the controls. A heavy growth was obtained from 7% of the patients’ swabs and 1 Yo of the swabs from the controls. Urinary and Introital Bacteria The results of the examination of introital swabs in the days preceding a bacteriuric episode were available in 28 of 30 episodes. Organisms considered to be similar when identified biochemically were present in the introitus and the urine in 23 of these 28 episodes. On 16 occasions E. coli was found in the introitus before appearing in the urine. On 12 occasions E. coli from both sources were serotyped and in only one case was the serotype different. The period of introital colonisation preceding a bacteriuric episode was variable. Eight of the episodes of bacteriuria followed a period of introital colonisation exceeding 5 days’ duration. Urinary and Faecal Bacteria For 28 bacteriuric episodes faecal specimens were obtained during the 6 days preceding the episode and for 11 of the episodes 2 specimens were obtained during this time. Similar organisms, when identified as described, were found in 19 of the 28 episodes. The urinary organism was E. coli in 16 of the episodes; serotyping was carried out in 11 of these and the similarity of the organisms
FAECAL AND INTROITAL BACTERIA AND URINARY TRACT INFECTION
was confirmed in 5. Micrococci were isolated from the urine on 6 occasions; they were never isolated from the faeces.
Haemolysin Production Eight per cent of isolates of E. coli from the faeces were haemolytic. There was no difference between the controls and the patients in the faecal carriage rate of haemolytic E. coli. Eighteen per cent of isolates from the introitus were haemolytic. This comprised 21% of the 1166 isolates from patients and 2% of the 150 isolates from the controls. No E. coli were isolated from the urine of the controls. Thirty-four per cent of the E. coli isolates from the urine of the patients were haemolytic.
319
obtained that introital colonisation may be of importance in the development of urinary tract infection. The results presented here support this view as introital colonisation was more frequent and heavier in the women with recurrent urinary infection that in the controls. In addition, the previous isolation of the infecting organism from the introitus was regularly demonstrated, as was the development of infection following a period of introital colonisation of more than 5 days’ duration. Although it is generally accepted that the bowel is the source of many of the organisms causing urinary infection, we were unable t o demonstrate in the faeces the serotype of E. coli which subsequently caused infection in 6 of the 11 patients in whom this was attempted. The reasons for this are not clear. Many of the investigations demonstrating faecal carriage of the infecting strain have been carried out subsequent t o the development of the infection (Gruneberg et a/., 1968; Bettelheim et al., 1971) and few prospective studies have been done. It has been shown that under certain conditions the faecal coliform population may change rapidly (Cooke et al., 1969) and although specimens of faeces were obtained twice-weekly it may be that the infecting strain was present in the faeces for such a short time before the development of the infection that it was not detected. The results for haemolysin production are similar to those obtained previously (Cooke and Ewins, 1975). The introital isolates from the patients were commonly haemolytic (2 19’0)whereas those from the controls were not (2070). Thirtyfour per cent of the urinary infecting strains were haemolytic. It has been shown (Cooke and Ewins, 1975) that introital strains that subsequently cause infection are more frequently haemolytic than those which d o not and the present findings support this view. This study of a small number of patients provides evidence that women with bacteriuria cannot be divided clearly into symptomatic and asymptomatic groups, as patients with symptoms also undergo asymptomatic episodes of bacteriuria and pyuria which outnumber the symptomatic episodes.
Discussion The initial aim of the investigation was to study in detail the relationship between introital colonisation and the development of urinary infection, but the finding which emerged most clearly was the frequency of asymptomatic infection in patients who were also suffering from recurrent symptomatic disease. Asymptomatic bacteriuria in non-pregnant women has been studied and it has been shown (Asscher et al., 1972) that symptomatic infection develops more frequently among bacteriuric than control subjects. In this investigation we have shown that in women with recurrent symptomatic infection the symptomatic episodes constitute less than a third of the episodes of bacteriuria and pyuria. Although symptoms were most frequently associated with bacteriuria and pyuria of more than 3 days’ duration and with pyuria of more than 60 cells/mm3, the possibility arises that these very frequent attacks of infection, which are untreated because they are asymptomatic, may be of importance in the ldevelopment of the disease. Radiological disease appears t o be more common in wholly asymptomatic patients (O’Grady el al., 1972). From the small number of patients studied here it appears that the distinction between asymptomatic and symptomatic patients may not be a clear one, as symptomatic patients also have many asymptomatic episodes. The relationship between introital colonisation and the development of urinary tract infection has Acknowledgements been studied by a number of workers (O’Grady We thank Mr L. Wilson for expert technical assistance, and et al., 1970; Stamey et al., 1971; Marsh et al., Dr K. Bettelheim and Mrs M . Chandler for typing the strains 1972; Stamey, 1973) and some evidence has been of E. coli.
320
BRITISH JOURNAL OF UROLOGY
Hinman, F. (1966).Mechanisms for the entry of bacteria and the establishment of urinary infection in female children. Asscher, A. W., Chick, S., Redford, N., Waters, W. E., Journal of Urology, 96, 546-550. Sussman, M., Evans, J. S., McLachlan, M. S. F. and Marsh, F. P., Murray, M. and Panchamia, P. (1972).The Williams, J. E. (1972). Natural history of asymptomatic relationship between bacterial cultures of the vaginal bacteriuria (ASB) in non-pregnant women. In Urinary introitus and urinary infection. British Journal of Urology, Tract Infection, ed. Brumfitt, W. and Asscher, A. W. 44,368-375. pp. 51-61.London: Oxford University Press. O’Crady, F. W., Charlton, C. A. C., Kelsey Fry, I., Balrd-Parker, A. G. (1966). Methods for classifying StaphylMcSherry, A. and Cattell, H’. R. (1972).Natural history of cocci and Micrococci. In Idenrification Methods for Microintractable ‘cystitis’ in women referred to a special clinic. biologists, ed. Gibbs, B. M. and Skinner, F. A. Pp. 59-64. In Urinary Tract Infedon, ed. Brumfitt, W.and Asscher, London: Academy Press. A. W . Pp. 81-91.London: Oxford University Press. Barnard, D., Bastable. J. R. G . and Richards, B. (1973). O’Grady, F. W., McSherry, M. A., Richards, B., O’Farreil, Estimation of residual urine volume with 1311-Hippuran. S. M. and Cattell, W. R. (1970). lntroital enterobacteria, British Journal of Urology, 45, 408-414. urinary infection and the urethral syndrome. Lancet, 2, Bettelheim, K. A., Dulake, C. and Taylor, J. (1971). Post1208-I210. operative urinary infections caused by Escherichia coli. Stamey, T. A. (1973). The role of introital enterobacteria Journal of Clinical Pathology, 24,442-443. in recurrent urinary infections. Journal of Urology, 109, Bettelheim, K. A. and Taylor, J. (1969). A study of Escher467-472. ichia coli isolated from chronic urinary infection. Journal Stamey, T. A., Timothy, M., Mlllar, M. and Mihara, G. of Medical Microbiology. 2, 225-236. (1971). Recurrent urinary infections in adult women: the Cooke, E. M. (1968). Properties of strains of Escherichia role of introital enterobacteria. California Medicine, 115, coli isolated from the faeces of patients with ulcerative 1-19. colitis, patients with acute diarrhoea and normal persons. Journal of Pathology and Bacteriology, 95, 101-113. Cooke, E. M. and Ewins, S. P. (1975).Properties of strains of Escherichia coli isolated from a variety of sources. The Authors Journal of Medical Microbiology, 0, 107-1I I. Cooke, E. M., Ewins, P. S. and Shooter, R. A. (1969). B. Richards, MD, FRCS, Consultant Urologist. Changing faecal population of Escherichia coli in hospital E. Mary Cooke, BSc, MD, MRCPath, Consultant Microbiologist, York. Now Professor of Clinical Microbiology, medical patients. British Medical Journal, 4,593-595. Institute of Pathology, Leeds. Gruneberg, R. N., Leigh, D. A. and Brumfitt, W. (1968). Escherichia coli serotypes in urinary tract infection: studies in domiciliary, ante-natal and hospital practice. In Urinary Tract Injection, ed. O’Grady, F. and Brumfitt, W. Requests for reprints to: B. Richards, Department of Urology, York District Hospital, Wigginton Road, York YO3 7HE. Pp. 68-69.London: Oxford University Press.
References
