175
mortality and morbidity and also reduced long-term neurological sequelae. This approach seems to be justified even in very early delivery (gestational age 27-30 wks) provided that adequate obstetric and neonatal intensive care is available. This study was supported by grants from the Prenatal Research Foundation (Prenatalforskningsnamnden), Malmohus County Council Research Foundation, and The Research Foundation "Margaretahemmet". -
Requests for reprints should be addressed to I.I., Department of ObsGynaecology, University of Lund, S-221 85 Lund, Sweden.
tetrics and
REFERENCES 1. 2.
Brander,
T. Finska LäkSällsk.
Årsskrift. 1936, 79, 603.
Rovinsky, J. J., Miller, J. A., Kaplan,
S.
Am. J.
Obstet.
Gynec. 1973, 115,
497. 3. Kauppila, O. Acta obstet. gynec. scand. 1975, suppl. 39. 4. Stewart, A. L., Reynolds, E. O.R. Pediatrics, 1974, 54, 724. 5. Kubli, F., Boos, W., Ruttgers, H. in Proceedings of the 5th European Congress of Perinatal Medicine (edited by G. Rooth and L. E. Bratteby); p. 69. Uppsala, 1976. 6. Goldenburg, R., Nelson, K. G.Am.J. Obstet. Gynec, 1977, 127, 240. 7. Cruikshank, D. P., Pitkin, R. M. Obstet, Gynec., N.Y., 1977, 50, 367. 8. Ingemarsson, I., Westgren, M., Svenningsen, N. W. Unpublished. 9. Ingemarsson, I.Am.J. Obstet. Gynec. 1976, 125, 520. 10. Liggins, G. C., Howie, R. N. Pediatrics, 1972, 50, 515. 11. Ahlström, H., Jonsson, B., Svenningsen, N. W. Archs Dis. Childh. 1976, 51, 13. 12. Carlsson, J., Svenningsen, N. W. Acta pædiat. scand. 1975, 64, 813. 13. Blennow, G., Svenningsen, N. W., Gustafsson, B., Sundén, B., Cronquist, S. Acta obstet. gynec. scand. 1977, 56, 189. 14. Low, J. A., Pancham, S. R., Worthington, D. Obstet. Gynec., N.Y., 1976, 15. 16.
47, 14. Milner, R. D. G. Br. J. Obstet. Gynœc. 1975, 82, 783. Crawford, J. S. ibid. 1974, 81, 867.
CONTROLLED TRIAL OF PROPHYLACTIC TREATMENT IN CHILDHOOD URINARY-TRACT INFECTION
J. M. SMELLIE
G. KATZ
R. N. GRÜNEBERG
University College Hospital, London, and Edgware General Hospital, Middlesex
Introduction IN children without obstruction to urinary outflow there is no difficulty in eradicating infection from the urinary tract with a short course of the appropriate antibacterial drug. The prevention of further infections, however, is more difficult. Infection recurs within 1-2 yrs in 50-75% of girls receiving a short course of treat-
mentu2 and 85-90% of such infections are reinfections with a fresh organism, not relapses.3 Long-term low-dosage prophylaxis is clinically effective and is widely used to prevent recurrent symptomatic infection in children,4-8 but few controlled trials have been carried out. Both co-trimoxazole and nitrofurantoin are effective prophylactic drugs.9-11 We report here a randomised controlled trial comparing low-dosage prophylaxis by co-trimoxazole and nitrofurantoin with no prophylaxis in a group of children with infection and radiologically normal urinary tracts, to determine (a) the efficacy of prophylaxis in preventing recurrent urinary infection and (b) whether long-term prophylaxis affected the pattern of recurrent infection after discontinuation of antibacterial therapy. Patients and Methods Patients 53 children aged 2-12 yrs with bacteriologically proven symptomatic urinary-tract infections and with normal intravenous urograms and micturating cystograms were admitted to the trial, with parental consent. 45 children (5 boys and 40 girls) completed the trial. 6 children failed to comply or attend regularly, and 2 children moved away.
Management The presenting infection was treated in 52 of the children by a short course (7-10 days) of co-trimoxazole, approximately 20 mg of sulphamethoxazole (s.M.x.) and 4 mg of trimethoprim (T.M.P.) /kg/day. 1 child received ampicillin 50 mg/kg/ day for 10 days instead of co-trimoxazole. Further management was randomised between no prophylaxis (50%), prophylactic nitrofurantoin (25%), and prophylactic co-trimoxazole (25%). The nature of the treatment was the clinician involved. Of the 8 children who did not the trial, 2 received no prophylaxis, 4 received nitrocomplete furantoin, and 2 co-trimoxazole. 2 of the 45 children, who initially received no prophylaxis and in whom recurrences developed after 5 and 7 mos, respectively, were readmitted to the trial. The randomisation process dictated that each received prophylactic co-trimoxazole, so that 47 episodes of infection were studied. The distribution of infection was as follows:1. 22 children (21 girls, 1 boy) received no further treatment and were followed-up regularly for 2 yrs, or until recurknown
In a randomised controlled trial 45 children (47 episodes of urinary-tract infection) with radiologically normal urinary tracts were given either low-dose prophylactic co-trimoxazole or nitrofurantoin or no prophylaxis after each had been treated with a short course of co-trimoxazole. During prophylaxis none of the 25 children had a further infection whereas half of the 22 who received no prophylaxis had a further infection within a similar period. 8 (32%) of those children who received prophylaxis had a further infection within 1 yr of completing the course of prophylactic treatment compared with 13 (59%) of the no-prophylaxis group in the year following the initial course of co-trimoxazole. 24 children had had a previous urinary-tract infection. All 6 of these who did not receive prophylaxis had a further infection whereas only 8 of the 18 (44%) who had received prophylaxis had a further infection after the course of prophylactic treatment. 24 mos after the start of the study, 68% of the prophylactic group and 36% of those who received no prophylaxis were recurrence-free. Differences in symptomatology and timing of reinfections and possible contributory factors are discussed.
Summary
to
rence.
2. 13 children (11 girls, 2 boys) received prophylactic lowdose co-trimoxazole, 10 mg S.M.X. and 2 mg T.M.p./kg/day, for 6-12 mos. All were followed-up for at least 1 yr after completing the course, and most for 2 yrs. 3. 12 children (10 girls, 2 boys) received prophylactic nitrofurantoin, 1-2 mg/kg/day, and were followed-up in the same way. Prophylaxis was given orally, either twice daily or in a single evening dose. The groups were similar in age and social-class distribution but a higher proportion of those receiving prophylaxis had a past history of symptomatic infection.
Bacteriological Methods to
Clean specimens of urine were collected, chilled immediately 4°C, and examined and plated on cysteine lactose electrolyte-
176 deficient (C.L.E.D.) agar within 1 h. Organisms were identified, rectal swabs processed, and compliance tested for at random by standard methods." Sensitivity tests were carried out for nitrofurantoin, sulphonamide, ampicillin, and trimethoprim.
of Treatment original plan was 6
Duration
The later extended mos’ prophylaxis. was
to
mos’ prophylaxis but the duration 1 yr. In all, 25 children received 243
Other Measures
Adequate fluid intake and regular bowel and voiding habits encouraged. Older children were instructed in double micturition at bedtime to ensure complete bladder emptying. were
Follow-up The children were seen 1 and 2 wks after diagnosis, and then at 1 mo, 3, 6, 9, and 12 mos both during and after the course of treatment and also if symptoms developed. Every child was observed for at least 2 yrs and those receiving prophylaxis, for 15 mos to 2 yrs after it was stopped. Urine specimens were obtained for culture at every visit and rectal swabs on most occasions.
Results
Presenting Infection The presenting infection
was due.to Escherichia coli. in all the 22 children who received only a short course of treatment and in 20 of the 25 children who received prophylaxis. Proteus mirabilis was isolated from 4 patients (3 of them boys) and Providentia spp. from another boy.
Recurrence
of Infection During Prophylaxis
of infection occurred in any of the 25 children who received prophylaxis during the time they were on their low-dosage co-trimoxazole or nitrofurantoin, a total of 243 mos (average 10 mos per child). In contrast, further infections occurred within the 10 mos after the initial course of treatment in 11 of the. 22 children who had no prophylaxis-in 7 (32%) of these 22 children the recurrence appeared within 6 mos, in 13 (59%) within 12 mos, and in 14 (64%) within 2 yrs. No
recurrence
Actuarial percentage recurrence-free curves over a minimum period of 24 mos in 22 children who did not receive prophylaxis and in 25 children who did.
Recurrence in Relation to Previous History of Urinary Infection No previous history.-23 children admitted to the trial had presented with their first known urinary-tract infection. None of the 7 children who received prophylaxis had a recurrent infection during the 12 mos after it was stopped whereas there was a further infection during the year of follow-up after initial treatment in 7 of the 16 who had no prophylaxis. (Fisher exact test pO.047) (table IB). History of previous infection.-During the year of follow-up of 24 children with a history of previous symp-
tomatic infection further infection occurred in all of the 6 who did not receive prophylaxis, but only 8 (44%) of the 18 who received prophylaxis had a recurrence within a year of stopping prophylactic treatment. (Fisher exact test
p=0.022) (table ic).
Recurrence of
Infection After Prophylaxis completing the course of low-dosage prophylaxis
Infection Over Two Years All children were seen regularly either until infection recurred (when they were eliminated from the trial) or
8 (32%) of the 25 children had a further infection within 1 yr compared with 13 (59%) of the 22 who had no prophylaxis (table IA). 5 of these infections occurred in the 13 children who had received co-trimoxazole and 3 in the 12 children who had received nitrofurantoin.
for at least 2 yrs. Children with a recurrence received 1 week’s full-dosage treatment with an appropriate drug. 2 of those who received co-trimoxazole were readmitted to the trial; the others were eliminated and managed according to our routine practice.
Recurrence of
After
TABLE I-RECURRENCES IN THE YEAR AFTER PROPHYLAXIS WAS STOPPED AND RECURRENCES IN THE YEAR AFTER A SHORT COURSE OF
TREATMENT BUT NO PROPHYLAXIS I
I
I
177
girl who had no prophylaxis had her first recurrence only in the second year, so that by 24 mos there had been recurrences in 14 (64%) of the 22 children who did not have prophylaxis and in 8 (32%) 1 further
TABLE II-RECURRENCE OF INFECTION DURING THE YEAR OF FOLLOW-UP IN THE TWO TREATMENT GROUPS IN RELATION TO POSSIBLE PREDISPOSING FACTORS
of the 25 children who did. Since children in this trial received prophylaxis for varying durations, actuarial percentage recurrence-free curves were compared by using the log rank test. At 2 yrs from the start of the study, there was a significantly greater recurrence-free rate amongst the children who received prophylaxis, (X2=6.99, d.f. 1, p=0.008) (see accompanying figure). 68% of the children who had received prophylaxis were recurrence-free compared with 36% of those who did not.
Clinical Features of
Infections Symptoms and timing of recurrences.-All the presenting infections were accompanied by symptoms, which included fever, dysuria, frequency, enuresis, abdominal pain, and hasmaturia. After prophylaxis, 7 of the 8 further infections were asymptomatic and were discovered only on routine urine culture; the other was associated with dysuria. They all occurred within 6 mos of stopping prophylaxis (follow-up of the 17 who did not
.
have a recurrence was continued for at least 1 yr and for 2 yrs in 11 of them). In contrast, only 1 of the recurrences in children who had no prophylaxis was asymptomatic ; 11 children complained of dysuria, frequency, or bed wetting, and 1 girl also had a high fever and general malaise. 7 of these 13 further infections occurred between 3 and 6 mos, and 6 between 7 and 12 mos, after the initial short course of treatment. One symptomatic infection appeared 19 mos after the initial short course of treatment. Associated factors in children who had further infections.-On entry to the trial, factors possibly predisposing to urinary-tract infection were identifiable from the clinical history or on investigation in 29 children. These included chronic constipation (7); low fluid intake and no drinks at school (3); infrequent voiding (three times or less daily) and not using school toilets (6), with a considerably enlarged bladder capacity in 3; a measurable bladder residue at cystography (14); and social problems (4). 20 (69%) of these children had a further infection-8 (57%) of the 14 who received prophylaxis and 12 (80%) of the 15 who did not (table n). Only 1 of the 18 children with no apparent predisposing factor on presentation had a recurrence. At the time of the recurrence, the main contributory factors had either persisted or returned and in more than half, the fresh infection and reversion to former voiding and drinking habits coincided with reduced parental supervision, during for-
eign
or
camping holidays
or
when there
was
parental
marital discord.
Microbiology Urinary organisms
and sensitivity.-20 of the recurinfections were caused by E. coli, 1 by Str. faecalis, and 1 was not identified. The antibacterial sensitivities of the organisms which caused the presenting urinary infections and the recurrences are shown in table ill. There was no significant difference, before and after treatment in the proportion of organisms which were fully sensitive to nitrofurantoin, ampicillin, sulphonamide, and trimethoprim. It was possible to distinguish between reinfection and relapse (or persistent infection) in 19 of the 22 recurrences. These 19 were reinfections and occurred in all 8 of the children who had received prophylaxis and in 11 of the 14 who had rent
not.
coliform identical with the organism at the time of infection was urinary organism in 22 of the 24 presenting infections and in all the 15 further infections for which this paired data was available. The results of culture of at least three serial rectal swabs were available from 35 children and showed that faecal coliforms were rapidly reduced or eliminated in 19 children after the full-dosage short course of co-trimoxazole. A fully sensitive coliform population returned within 2-4 wks in children receiving prophylactic nitrofurantoin or no prophylaxis. The effect on faecal coliforms continued during prophylaxis in 8 children receiving co-trimoxazole but the bowel flora rapidly returned to normal after co-trimoxazole was withdrawn. No trimethoprim-resistant organisms emerged either in the bowel flora or in the urine during this study. Rectal
flora.-The predominant rectal
Discussion This controlled trial shows that while it is
TABLE III-ANTIBACTERIAL SENSITIVITY PATTERN OF THE URINARY PATHOGEN AT PRESENTATION AND AT RECURRENCE
Sulph.=sulphonamide; Amp.=ampicillin; N.F.=nitrofurantoin; N.K.=sensitivity of organism not known.
being
178
given, long-term low-dosage prophylaxis prevents further infection of the unobstructed urinary tract. This is in keeping with previous uncontrolled studies of prophylaxis and a recent controlled trial of nitrofurantoin.12 In the trial reported here, only children with symptomatic infection of radiologically normal urinary tracts were studied but the value of low-dosage prophylaxis has previously been demonstrated in uncontrolled studies of children with vesico-ureteric reflux, renal scarring, or morphological anomalies such as duplex kid-
ney.11,13—15
REFERENCES 1.
Kunin,
C.
M., Deutscher, R., Paquin,
91.
It must be
emphasised that low-dosage prophylaxis will not prevent persisting or recurrent infection if there is continuing urinary stasis, or obstruction to urinary outflow which needs surgical relief. The established relation between an increased residual urine and recurrence of infection 16, 17 and the association of constipation and urinary infection,’8 suggest that regular complete bladder emptying and the prevention of faecal overloading are important in preventing further infection. There was no significant difference in the incidence of recurrence in the year after only a short course of antibacterial treatment and in that following prophylaxis, but when the previous history of infection was taken into account, the observed differences in recurrence in the two groups was significant. Prophylaxis may have provided time for the inflamed bladder mucosa to recover or for the defect in bladder defence to be corrected. These results, together with the finding that at least 90% of the further infections were reinfections, support the view that the main cause of recurrent infection in the absence of outflow obstruction is a breakdown of the host defences,19,20 allowing colonisation of the susceptible bladder by bowel commensals ascending the urethra. It follows that the effect upon the bowel flora of different antibacterial drugs and of the dosage used to eradicate the presenting infection or in subsequent prophylaxis is of considerable importance, especially in determining the nature and sensitivity of a further infection. In all the recurrences tested the predominant faecal organism was identical with the urinary organism so that a rectal swab may be of predictive therapeutic value. As in our earlier studies,21,22 we found an almost immediate reduction in the bowel coliform population both after full-dosage and during low-dosage co-trimoxazole treatment, with a rapid return to normal counts of sensitive organisms on its withdrawal. No trimethoprim or nitrofurantoin resistance developed during this study. There was no increase, during prophylaxis, of the level of sulphonamide-resistance normally found and this was evenly distributed between the groups. The difference between the two groups in symptomatology and timing of reinfections is of some interest in view of recent reports of differences in the strains of organisms causing symptomatic and asymptomatic bacteriuria.23,24 Clearly low-dosage antibacterial prophylaxis with cotrimoxazole or nitrofurantoin prevents recurrence of urinary infection in the unobstructed urinary tract. Its principal use at present is to control recurrent symptomatic infections and to prevent infection in infants and children with vesicoureteric reflux. We thank Dr Peter
for technical microbiological assistance; Sister March and staff of the Children’s Department, University College Hospital for their help; local general practitioners for referring untreated children for diagnosis ; Mrs Marjorie Bedford for secretarial help; and Mr P. Fayers and Dr A. Johnson of the Medical Research Council Statistical Research Unit. This study was supported by a grant from the Medical Research Council. Requests for reprints should be addressed to J.M.S., Paediatric Department, University College Hospital, Huntley Street, London WC1E 6AU
Wimberley for help in the clinics; Mrs A. Leakey
A.
Medicine, (Baltimore), 1964, 43,
2.
Bergström, T., Lincoln, K., Redin, B., Winberg, J. Acta pœdiat. scand. 1968, 57, 186. 3. Grüneberg, R. N., Smellie, J. M., Leakey, A. in Urinary Tract Infection II, (edited by W. Brumfitt and A. W. Asscher); p. 131. London, 1973. 4. Normand, I. C. S., Smellie, J. M. Br. med. J. 1965, i, 1023. 5. MacGregor, M., Freeman, P. Q.Jl Med. 1975, 44, 481. 6. White, R. H. R. Br. med. J. 1977, i, 1650. 7. King, L. R., Kazmi, S. O., Belman, A. B. Urol. Clin. N. Am. 1974, 1, 441. 8. Govan, D. E., Fair, W. R., Friedland, G. W., Filly, R. A. Urology, 1975, 6, 273. 9.
O’Grady, F., Fry, I. K., McSherry, A., Cattell, W. R. J. infect. 128, (suppl), S652. 10. Bailey, R. R., Gower, P. E., Roberts, A. P., de Wardener, H.
Dis.
1973,
P. Lancet, 1971, ii, 1112. 11. Smellie, J. M., Grüneberg, R. N., Leakey, A., Atkin, W. S. Br. med. J. 1978, ii, 203. 12. Lohr, J. A., Nunley, D. H., Howards, S. S., Ford, R. F. Pediatrics, Springfield, 1977, 59, 562. 13. Edwards, D., Normand, I. C. S., Prescod, N., Smellie, J. M. Br. med. J. 1977, ii, 285. 14. Forbes, P. A., Drummond, K. N.J. infect. Dis. 1973, 128, (suppl), S626. 15. MacGregor, M. E., Wynne Williams, C. J. E. Lancet, 1966, i, 893 16. Shand, D. G., Nimmon, C. C., O’Grady, F., Cattell, W. R. ibid. 1970, i,
1305.
Lindberg, V., Bjure, J., Haugstvedt, S., Jodal, U. Acta pœdiat. scand. 1975, 64, 437. 18. Neumann, P. Z., de Domenico, I. J., Nogrady, M. B. Pediatrics, Springfield, 1973, 52, 241. 19. O’Grady, F., Cattell, W. R. Br. J. Urol. 1966, 38, 156. 20. Stamey, T. A., Condy, M., Milhara, G. New Engl. J. Med. 1977, 296, 780. 21. Grüneberg, R. N., Smellie, J. M., Leakey, A., Atkin, W. S. Br. med. J. 1976, 17.
ii, 206. 22. 23.
Grüneberg, R. N., Leakey, A., Bendall, M. J., Smellie, J. M. Kidney Int. 1975, 8, (suppl 4), S122. Lindberg, V., Hanson, L. A., Jodal, U., Lidin-Janson, G., Lincoln, K., Olling, S. Acta pœdiat. scand. 1975, 64, 432. J. D., Waterman, A. M., Gower, P. E., Koutsaimanis, K.G.J. med. Microbiol. 1975, 8, 311.
24. Roberts, A. P., Linton,
FACTORS INFLUENCING LACTATION PERFORMANCE IN RURAL GAMBIAN MOTHERS R. G. WHITEHEAD MELANIE HUTTON
M. G. M. ROWLAND
ELISABETH MÜLLER
ALISON PAUL
A. M. PRENTICE
Medical Research Council Dunn Nutrition Unit, Milton Road, Cambridge CB4 1XJ, and Keneba, The Gambia.
Breast-milk consumption has been measured in a rural African community in which breast-feeding on demand is universally practised until the baby is 18 mos old. The mother’s long-term capacity for breast-milk production is determined by the end of the second month of lactation, yield being closely correlated with the infant’s birth-weight. Other factors significantly influencing output were parity, month of lactation, baby’s weight-for-age, season, and maternal diet. Daily milk consumption was limited primarily by the amount delivered per feed, not the frequency of feed-
Summary
ing. THE
Introduction universal shortage of information
on
the
mortality and morbidity and also reduced long-term neurological sequelae. This approach seems to be justified even in very early delivery (gestational age 27-30 wks) provided that adequate obstetric and neonatal intensive care is available. This study was supported by grants from the Prenatal Research Foundation (Prenatalforskningsnamnden), Malmohus County Council Research Foundation, and The Research Foundation "Margaretahemmet". -
Requests for reprints should be addressed to I.I., Department of ObsGynaecology, University of Lund, S-221 85 Lund, Sweden.
tetrics and
REFERENCES 1. 2.
Brander,
T. Finska LäkSällsk.
Årsskrift. 1936, 79, 603.
Rovinsky, J. J., Miller, J. A., Kaplan,
S.
Am. J.
Obstet.
Gynec. 1973, 115,
497. 3. Kauppila, O. Acta obstet. gynec. scand. 1975, suppl. 39. 4. Stewart, A. L., Reynolds, E. O.R. Pediatrics, 1974, 54, 724. 5. Kubli, F., Boos, W., Ruttgers, H. in Proceedings of the 5th European Congress of Perinatal Medicine (edited by G. Rooth and L. E. Bratteby); p. 69. Uppsala, 1976. 6. Goldenburg, R., Nelson, K. G.Am.J. Obstet. Gynec, 1977, 127, 240. 7. Cruikshank, D. P., Pitkin, R. M. Obstet, Gynec., N.Y., 1977, 50, 367. 8. Ingemarsson, I., Westgren, M., Svenningsen, N. W. Unpublished. 9. Ingemarsson, I.Am.J. Obstet. Gynec. 1976, 125, 520. 10. Liggins, G. C., Howie, R. N. Pediatrics, 1972, 50, 515. 11. Ahlström, H., Jonsson, B., Svenningsen, N. W. Archs Dis. Childh. 1976, 51, 13. 12. Carlsson, J., Svenningsen, N. W. Acta pædiat. scand. 1975, 64, 813. 13. Blennow, G., Svenningsen, N. W., Gustafsson, B., Sundén, B., Cronquist, S. Acta obstet. gynec. scand. 1977, 56, 189. 14. Low, J. A., Pancham, S. R., Worthington, D. Obstet. Gynec., N.Y., 1976, 15. 16.
47, 14. Milner, R. D. G. Br. J. Obstet. Gynœc. 1975, 82, 783. Crawford, J. S. ibid. 1974, 81, 867.
CONTROLLED TRIAL OF PROPHYLACTIC TREATMENT IN CHILDHOOD URINARY-TRACT INFECTION
J. M. SMELLIE
G. KATZ
R. N. GRÜNEBERG
University College Hospital, London, and Edgware General Hospital, Middlesex
Introduction IN children without obstruction to urinary outflow there is no difficulty in eradicating infection from the urinary tract with a short course of the appropriate antibacterial drug. The prevention of further infections, however, is more difficult. Infection recurs within 1-2 yrs in 50-75% of girls receiving a short course of treat-
mentu2 and 85-90% of such infections are reinfections with a fresh organism, not relapses.3 Long-term low-dosage prophylaxis is clinically effective and is widely used to prevent recurrent symptomatic infection in children,4-8 but few controlled trials have been carried out. Both co-trimoxazole and nitrofurantoin are effective prophylactic drugs.9-11 We report here a randomised controlled trial comparing low-dosage prophylaxis by co-trimoxazole and nitrofurantoin with no prophylaxis in a group of children with infection and radiologically normal urinary tracts, to determine (a) the efficacy of prophylaxis in preventing recurrent urinary infection and (b) whether long-term prophylaxis affected the pattern of recurrent infection after discontinuation of antibacterial therapy. Patients and Methods Patients 53 children aged 2-12 yrs with bacteriologically proven symptomatic urinary-tract infections and with normal intravenous urograms and micturating cystograms were admitted to the trial, with parental consent. 45 children (5 boys and 40 girls) completed the trial. 6 children failed to comply or attend regularly, and 2 children moved away.
Management The presenting infection was treated in 52 of the children by a short course (7-10 days) of co-trimoxazole, approximately 20 mg of sulphamethoxazole (s.M.x.) and 4 mg of trimethoprim (T.M.P.) /kg/day. 1 child received ampicillin 50 mg/kg/ day for 10 days instead of co-trimoxazole. Further management was randomised between no prophylaxis (50%), prophylactic nitrofurantoin (25%), and prophylactic co-trimoxazole (25%). The nature of the treatment was the clinician involved. Of the 8 children who did not the trial, 2 received no prophylaxis, 4 received nitrocomplete furantoin, and 2 co-trimoxazole. 2 of the 45 children, who initially received no prophylaxis and in whom recurrences developed after 5 and 7 mos, respectively, were readmitted to the trial. The randomisation process dictated that each received prophylactic co-trimoxazole, so that 47 episodes of infection were studied. The distribution of infection was as follows:1. 22 children (21 girls, 1 boy) received no further treatment and were followed-up regularly for 2 yrs, or until recurknown
In a randomised controlled trial 45 children (47 episodes of urinary-tract infection) with radiologically normal urinary tracts were given either low-dose prophylactic co-trimoxazole or nitrofurantoin or no prophylaxis after each had been treated with a short course of co-trimoxazole. During prophylaxis none of the 25 children had a further infection whereas half of the 22 who received no prophylaxis had a further infection within a similar period. 8 (32%) of those children who received prophylaxis had a further infection within 1 yr of completing the course of prophylactic treatment compared with 13 (59%) of the no-prophylaxis group in the year following the initial course of co-trimoxazole. 24 children had had a previous urinary-tract infection. All 6 of these who did not receive prophylaxis had a further infection whereas only 8 of the 18 (44%) who had received prophylaxis had a further infection after the course of prophylactic treatment. 24 mos after the start of the study, 68% of the prophylactic group and 36% of those who received no prophylaxis were recurrence-free. Differences in symptomatology and timing of reinfections and possible contributory factors are discussed.
Summary
to
rence.
2. 13 children (11 girls, 2 boys) received prophylactic lowdose co-trimoxazole, 10 mg S.M.X. and 2 mg T.M.p./kg/day, for 6-12 mos. All were followed-up for at least 1 yr after completing the course, and most for 2 yrs. 3. 12 children (10 girls, 2 boys) received prophylactic nitrofurantoin, 1-2 mg/kg/day, and were followed-up in the same way. Prophylaxis was given orally, either twice daily or in a single evening dose. The groups were similar in age and social-class distribution but a higher proportion of those receiving prophylaxis had a past history of symptomatic infection.
Bacteriological Methods to
Clean specimens of urine were collected, chilled immediately 4°C, and examined and plated on cysteine lactose electrolyte-
176 deficient (C.L.E.D.) agar within 1 h. Organisms were identified, rectal swabs processed, and compliance tested for at random by standard methods." Sensitivity tests were carried out for nitrofurantoin, sulphonamide, ampicillin, and trimethoprim.
of Treatment original plan was 6
Duration
The later extended mos’ prophylaxis. was
to
mos’ prophylaxis but the duration 1 yr. In all, 25 children received 243
Other Measures
Adequate fluid intake and regular bowel and voiding habits encouraged. Older children were instructed in double micturition at bedtime to ensure complete bladder emptying. were
Follow-up The children were seen 1 and 2 wks after diagnosis, and then at 1 mo, 3, 6, 9, and 12 mos both during and after the course of treatment and also if symptoms developed. Every child was observed for at least 2 yrs and those receiving prophylaxis, for 15 mos to 2 yrs after it was stopped. Urine specimens were obtained for culture at every visit and rectal swabs on most occasions.
Results
Presenting Infection The presenting infection
was due.to Escherichia coli. in all the 22 children who received only a short course of treatment and in 20 of the 25 children who received prophylaxis. Proteus mirabilis was isolated from 4 patients (3 of them boys) and Providentia spp. from another boy.
Recurrence
of Infection During Prophylaxis
of infection occurred in any of the 25 children who received prophylaxis during the time they were on their low-dosage co-trimoxazole or nitrofurantoin, a total of 243 mos (average 10 mos per child). In contrast, further infections occurred within the 10 mos after the initial course of treatment in 11 of the. 22 children who had no prophylaxis-in 7 (32%) of these 22 children the recurrence appeared within 6 mos, in 13 (59%) within 12 mos, and in 14 (64%) within 2 yrs. No
recurrence
Actuarial percentage recurrence-free curves over a minimum period of 24 mos in 22 children who did not receive prophylaxis and in 25 children who did.
Recurrence in Relation to Previous History of Urinary Infection No previous history.-23 children admitted to the trial had presented with their first known urinary-tract infection. None of the 7 children who received prophylaxis had a recurrent infection during the 12 mos after it was stopped whereas there was a further infection during the year of follow-up after initial treatment in 7 of the 16 who had no prophylaxis. (Fisher exact test pO.047) (table IB). History of previous infection.-During the year of follow-up of 24 children with a history of previous symp-
tomatic infection further infection occurred in all of the 6 who did not receive prophylaxis, but only 8 (44%) of the 18 who received prophylaxis had a recurrence within a year of stopping prophylactic treatment. (Fisher exact test
p=0.022) (table ic).
Recurrence of
Infection After Prophylaxis completing the course of low-dosage prophylaxis
Infection Over Two Years All children were seen regularly either until infection recurred (when they were eliminated from the trial) or
8 (32%) of the 25 children had a further infection within 1 yr compared with 13 (59%) of the 22 who had no prophylaxis (table IA). 5 of these infections occurred in the 13 children who had received co-trimoxazole and 3 in the 12 children who had received nitrofurantoin.
for at least 2 yrs. Children with a recurrence received 1 week’s full-dosage treatment with an appropriate drug. 2 of those who received co-trimoxazole were readmitted to the trial; the others were eliminated and managed according to our routine practice.
Recurrence of
After
TABLE I-RECURRENCES IN THE YEAR AFTER PROPHYLAXIS WAS STOPPED AND RECURRENCES IN THE YEAR AFTER A SHORT COURSE OF
TREATMENT BUT NO PROPHYLAXIS I
I
I
177
girl who had no prophylaxis had her first recurrence only in the second year, so that by 24 mos there had been recurrences in 14 (64%) of the 22 children who did not have prophylaxis and in 8 (32%) 1 further
TABLE II-RECURRENCE OF INFECTION DURING THE YEAR OF FOLLOW-UP IN THE TWO TREATMENT GROUPS IN RELATION TO POSSIBLE PREDISPOSING FACTORS
of the 25 children who did. Since children in this trial received prophylaxis for varying durations, actuarial percentage recurrence-free curves were compared by using the log rank test. At 2 yrs from the start of the study, there was a significantly greater recurrence-free rate amongst the children who received prophylaxis, (X2=6.99, d.f. 1, p=0.008) (see accompanying figure). 68% of the children who had received prophylaxis were recurrence-free compared with 36% of those who did not.
Clinical Features of
Infections Symptoms and timing of recurrences.-All the presenting infections were accompanied by symptoms, which included fever, dysuria, frequency, enuresis, abdominal pain, and hasmaturia. After prophylaxis, 7 of the 8 further infections were asymptomatic and were discovered only on routine urine culture; the other was associated with dysuria. They all occurred within 6 mos of stopping prophylaxis (follow-up of the 17 who did not
.
have a recurrence was continued for at least 1 yr and for 2 yrs in 11 of them). In contrast, only 1 of the recurrences in children who had no prophylaxis was asymptomatic ; 11 children complained of dysuria, frequency, or bed wetting, and 1 girl also had a high fever and general malaise. 7 of these 13 further infections occurred between 3 and 6 mos, and 6 between 7 and 12 mos, after the initial short course of treatment. One symptomatic infection appeared 19 mos after the initial short course of treatment. Associated factors in children who had further infections.-On entry to the trial, factors possibly predisposing to urinary-tract infection were identifiable from the clinical history or on investigation in 29 children. These included chronic constipation (7); low fluid intake and no drinks at school (3); infrequent voiding (three times or less daily) and not using school toilets (6), with a considerably enlarged bladder capacity in 3; a measurable bladder residue at cystography (14); and social problems (4). 20 (69%) of these children had a further infection-8 (57%) of the 14 who received prophylaxis and 12 (80%) of the 15 who did not (table n). Only 1 of the 18 children with no apparent predisposing factor on presentation had a recurrence. At the time of the recurrence, the main contributory factors had either persisted or returned and in more than half, the fresh infection and reversion to former voiding and drinking habits coincided with reduced parental supervision, during for-
eign
or
camping holidays
or
when there
was
parental
marital discord.
Microbiology Urinary organisms
and sensitivity.-20 of the recurinfections were caused by E. coli, 1 by Str. faecalis, and 1 was not identified. The antibacterial sensitivities of the organisms which caused the presenting urinary infections and the recurrences are shown in table ill. There was no significant difference, before and after treatment in the proportion of organisms which were fully sensitive to nitrofurantoin, ampicillin, sulphonamide, and trimethoprim. It was possible to distinguish between reinfection and relapse (or persistent infection) in 19 of the 22 recurrences. These 19 were reinfections and occurred in all 8 of the children who had received prophylaxis and in 11 of the 14 who had rent
not.
coliform identical with the organism at the time of infection was urinary organism in 22 of the 24 presenting infections and in all the 15 further infections for which this paired data was available. The results of culture of at least three serial rectal swabs were available from 35 children and showed that faecal coliforms were rapidly reduced or eliminated in 19 children after the full-dosage short course of co-trimoxazole. A fully sensitive coliform population returned within 2-4 wks in children receiving prophylactic nitrofurantoin or no prophylaxis. The effect on faecal coliforms continued during prophylaxis in 8 children receiving co-trimoxazole but the bowel flora rapidly returned to normal after co-trimoxazole was withdrawn. No trimethoprim-resistant organisms emerged either in the bowel flora or in the urine during this study. Rectal
flora.-The predominant rectal
Discussion This controlled trial shows that while it is
TABLE III-ANTIBACTERIAL SENSITIVITY PATTERN OF THE URINARY PATHOGEN AT PRESENTATION AND AT RECURRENCE
Sulph.=sulphonamide; Amp.=ampicillin; N.F.=nitrofurantoin; N.K.=sensitivity of organism not known.
being
178
given, long-term low-dosage prophylaxis prevents further infection of the unobstructed urinary tract. This is in keeping with previous uncontrolled studies of prophylaxis and a recent controlled trial of nitrofurantoin.12 In the trial reported here, only children with symptomatic infection of radiologically normal urinary tracts were studied but the value of low-dosage prophylaxis has previously been demonstrated in uncontrolled studies of children with vesico-ureteric reflux, renal scarring, or morphological anomalies such as duplex kid-
ney.11,13—15
REFERENCES 1.
Kunin,
C.
M., Deutscher, R., Paquin,
91.
It must be
emphasised that low-dosage prophylaxis will not prevent persisting or recurrent infection if there is continuing urinary stasis, or obstruction to urinary outflow which needs surgical relief. The established relation between an increased residual urine and recurrence of infection 16, 17 and the association of constipation and urinary infection,’8 suggest that regular complete bladder emptying and the prevention of faecal overloading are important in preventing further infection. There was no significant difference in the incidence of recurrence in the year after only a short course of antibacterial treatment and in that following prophylaxis, but when the previous history of infection was taken into account, the observed differences in recurrence in the two groups was significant. Prophylaxis may have provided time for the inflamed bladder mucosa to recover or for the defect in bladder defence to be corrected. These results, together with the finding that at least 90% of the further infections were reinfections, support the view that the main cause of recurrent infection in the absence of outflow obstruction is a breakdown of the host defences,19,20 allowing colonisation of the susceptible bladder by bowel commensals ascending the urethra. It follows that the effect upon the bowel flora of different antibacterial drugs and of the dosage used to eradicate the presenting infection or in subsequent prophylaxis is of considerable importance, especially in determining the nature and sensitivity of a further infection. In all the recurrences tested the predominant faecal organism was identical with the urinary organism so that a rectal swab may be of predictive therapeutic value. As in our earlier studies,21,22 we found an almost immediate reduction in the bowel coliform population both after full-dosage and during low-dosage co-trimoxazole treatment, with a rapid return to normal counts of sensitive organisms on its withdrawal. No trimethoprim or nitrofurantoin resistance developed during this study. There was no increase, during prophylaxis, of the level of sulphonamide-resistance normally found and this was evenly distributed between the groups. The difference between the two groups in symptomatology and timing of reinfections is of some interest in view of recent reports of differences in the strains of organisms causing symptomatic and asymptomatic bacteriuria.23,24 Clearly low-dosage antibacterial prophylaxis with cotrimoxazole or nitrofurantoin prevents recurrence of urinary infection in the unobstructed urinary tract. Its principal use at present is to control recurrent symptomatic infections and to prevent infection in infants and children with vesicoureteric reflux. We thank Dr Peter
for technical microbiological assistance; Sister March and staff of the Children’s Department, University College Hospital for their help; local general practitioners for referring untreated children for diagnosis ; Mrs Marjorie Bedford for secretarial help; and Mr P. Fayers and Dr A. Johnson of the Medical Research Council Statistical Research Unit. This study was supported by a grant from the Medical Research Council. Requests for reprints should be addressed to J.M.S., Paediatric Department, University College Hospital, Huntley Street, London WC1E 6AU
Wimberley for help in the clinics; Mrs A. Leakey
A.
Medicine, (Baltimore), 1964, 43,
2.
Bergström, T., Lincoln, K., Redin, B., Winberg, J. Acta pœdiat. scand. 1968, 57, 186. 3. Grüneberg, R. N., Smellie, J. M., Leakey, A. in Urinary Tract Infection II, (edited by W. Brumfitt and A. W. Asscher); p. 131. London, 1973. 4. Normand, I. C. S., Smellie, J. M. Br. med. J. 1965, i, 1023. 5. MacGregor, M., Freeman, P. Q.Jl Med. 1975, 44, 481. 6. White, R. H. R. Br. med. J. 1977, i, 1650. 7. King, L. R., Kazmi, S. O., Belman, A. B. Urol. Clin. N. Am. 1974, 1, 441. 8. Govan, D. E., Fair, W. R., Friedland, G. W., Filly, R. A. Urology, 1975, 6, 273. 9.
O’Grady, F., Fry, I. K., McSherry, A., Cattell, W. R. J. infect. 128, (suppl), S652. 10. Bailey, R. R., Gower, P. E., Roberts, A. P., de Wardener, H.
Dis.
1973,
P. Lancet, 1971, ii, 1112. 11. Smellie, J. M., Grüneberg, R. N., Leakey, A., Atkin, W. S. Br. med. J. 1978, ii, 203. 12. Lohr, J. A., Nunley, D. H., Howards, S. S., Ford, R. F. Pediatrics, Springfield, 1977, 59, 562. 13. Edwards, D., Normand, I. C. S., Prescod, N., Smellie, J. M. Br. med. J. 1977, ii, 285. 14. Forbes, P. A., Drummond, K. N.J. infect. Dis. 1973, 128, (suppl), S626. 15. MacGregor, M. E., Wynne Williams, C. J. E. Lancet, 1966, i, 893 16. Shand, D. G., Nimmon, C. C., O’Grady, F., Cattell, W. R. ibid. 1970, i,
1305.
Lindberg, V., Bjure, J., Haugstvedt, S., Jodal, U. Acta pœdiat. scand. 1975, 64, 437. 18. Neumann, P. Z., de Domenico, I. J., Nogrady, M. B. Pediatrics, Springfield, 1973, 52, 241. 19. O’Grady, F., Cattell, W. R. Br. J. Urol. 1966, 38, 156. 20. Stamey, T. A., Condy, M., Milhara, G. New Engl. J. Med. 1977, 296, 780. 21. Grüneberg, R. N., Smellie, J. M., Leakey, A., Atkin, W. S. Br. med. J. 1976, 17.
ii, 206. 22. 23.
Grüneberg, R. N., Leakey, A., Bendall, M. J., Smellie, J. M. Kidney Int. 1975, 8, (suppl 4), S122. Lindberg, V., Hanson, L. A., Jodal, U., Lidin-Janson, G., Lincoln, K., Olling, S. Acta pœdiat. scand. 1975, 64, 432. J. D., Waterman, A. M., Gower, P. E., Koutsaimanis, K.G.J. med. Microbiol. 1975, 8, 311.
24. Roberts, A. P., Linton,
FACTORS INFLUENCING LACTATION PERFORMANCE IN RURAL GAMBIAN MOTHERS R. G. WHITEHEAD MELANIE HUTTON
M. G. M. ROWLAND
ELISABETH MÜLLER
ALISON PAUL
A. M. PRENTICE
Medical Research Council Dunn Nutrition Unit, Milton Road, Cambridge CB4 1XJ, and Keneba, The Gambia.
Breast-milk consumption has been measured in a rural African community in which breast-feeding on demand is universally practised until the baby is 18 mos old. The mother’s long-term capacity for breast-milk production is determined by the end of the second month of lactation, yield being closely correlated with the infant’s birth-weight. Other factors significantly influencing output were parity, month of lactation, baby’s weight-for-age, season, and maternal diet. Daily milk consumption was limited primarily by the amount delivered per feed, not the frequency of feed-
Summary
ing. THE
Introduction universal shortage of information
on
the
