Special aspects in the diagnosis and treatment of urinary tract infections in pregnancy and the puerperium

Special aspects in the diagnosis and treatment of urinary tract infections in pregnancy and the puerperium

Special aspects in the diagnosis and treatment of urinary tract infections in pregnancy and the puerperium M. F. VICCHI, M.D. Quita, Ecuador Catheteri...

517KB Sizes 0 Downloads 107 Views

Special aspects in the diagnosis and treatment of urinary tract infections in pregnancy and the puerperium M. F. VICCHI, M.D. Quita, Ecuador Catheterization specimens of urine are more suitable for bacteriologic examination than clean-voided midstream specimens, but .catheJerization int;Tea:res the likelihood be of bacteriuria in subsequent examinations, This iatrogenic contamination avoided by administration of antibacterial agents, the most appropriate being nitrofurazone, by intravesical instillations. Positive results of bat:teriologic ur 'ne culture could be predicted in over 90 per cent of specimens by examination of uncentrifuged Gram-stained urine smears, This technique allows e4rlier effective treatment in urinary tract infection. Nitrofurantoin, preferably in the form of sodium salt administered intravenously, was more effective than combined demethylchlortetracycline-sulfamethoxypyridazine in the medical management of urinary tract infection in 101 pregnant women.

·can

U R 1 N A R Y tract infection in pregnancy and the puerperium has had increasing attention from obstetricians and internists the world over. Until recent years pyuria and bacteriuria in obstetric patients were accepted somewhat fatalistically as due to unavoidable mechanical and endocrine factors. But this attitude has been yielding to evidence that even asymptomatic bacteriuria during pregnancy, if untreated, may progress to overt clinical infection, chronic pyelonephritis, and hypertension. 1 - 4 Studies have indicated that urinary tract infection during pregnancy is associated also with increased incidence of prematurity, neonatal morbidity, and mortaiity.s-s These views are by no means universally accepted by clinicians. Also there is no consensus as to differential diagnostic criteria of infection, the optimal mode of therapy, or criteria of cure. Major studies published in the United States, Latin America, and Europe show disagreement on basic principles. The studies reported below were de-

signed to seek optimal methods of diagnosis and treatment in urinary tract infection during pr~ancy and the puerperium; under conditions which prevail in Latin American urban obstetric outpatient and maternity hospital practice. Over the past 6 years we have examined and treated nearly 600- obstetric patients for urinary tract infection, most of them at the Maternity Hospital Isidro de Ayora in Quinto. This 450 bed hospital draws its patients from the lower middle socioeconomic strata of the capital. Most of the patients are of Indian or mixed racial origin, and the educational level is low. Although this population is in the eariy stages of rapid industrialization and urbanizati
From the Hospital de Maternidad Isidro de Ayora

994

Urinary tract infections

Volume 99 Number 7

requirements for expensive equipment and highly skilled technical aides; ( 2) the need for obtaining specimens of urine by catheter, and the dangers of introducing infection by catheterization; and ( 3) the effectiveness of antibacterial prophylaxis and therapy. Laboratory diagnosis

Methods. We correlated the results of semiquantitative bacterial culture of the urine in 400 obstetric patients with results of immediate Gram stain on uncentrifuged specimens of urine. When specimens were not processed immediately they were refrigerated at 5° C.; then 0.2 mi. of urine was mixed with 7.8 ml. of broth (1 :40 dilution) and a drop (0.04 ml.) of the dilution was plated uniformly on trypticase soy broth and chloride-deficient media (for identification of Proteus species) in test tubes and Petri dishes, and incubated for 24 to 48 hours. Results. The finding of bacteria in a Gramstained, uncentrifuged specimen predicts the results of culture, in a great preponderance of instances (Table I). If the finding of 100,000 or more bacteria per milliliter of

995

Turbidity appeared within the first 5 hours of incubation only in those specimens which contained over 100,000 microorganisms per milliliter, while lack of visible growth at the end of the eighth incubation hour was correlated always with counts lower than 5,000 colonies per milliliter. Correct diagnosis of infection, permitting early effective treatment, in many instances could be postulated during the first 24 hours. Sternheimer-Malbin staining9 of urinary sediment was positive in 11 of 29 cases, all of which had over 100,000 organisms per milliliter of urine. This procedure, in which an alcoholic gentian violet-safranin solution is used to stain leukocytes, in our hands yielded too many false negative results to be used as a routine screening for pyelonephritis. It may aid in clinical diagnosis of this condition, however, when positive results are reported. It may be noted that urinary tract infection was present in 84 (21 per cent) of the 400 pregnant or puerperal women in this study, according to the culture criterion of 10,000 or more organisms per milliliter of urine.

urine is taken as the sen1iquantitative culture

criterion of infection, our data show that infection could be diagnosed in 97 per cent (387 of 400) of the cases by inspection of the immediate stained smear of uncentrifuged urine. If a lower level of bacteriuria, 10,000 bacteria per milliliter of urine, is taken as the cuiture criterion of infection, laboratory diagnosis by inspection of the stained smear predicted accurately in 94 per cent (376 of 400) of the cases. It was noted that bacterial counts could be predicted, roughly but reliably, by periodic inspection of culture tubes during incubation.

Catheterization

Methods. In order to evaluate the diagnostic adequacy of urine specimens obtained without catheterization we carried out semiquantitative cultures of urine on 30 obstetric patients. Two specimens were collected from each patient. The first was a dean-voided midstream specimen obtained after scrupulous cleaning and disinfection of the vestibule

and urethral meatus area. The second was collected on the following day, by catheterization with the usual precautions. In 213 other pregnant women in whom a

Table I. Correlation of Gram smear of uncentrifuged urine with semiquantitative urine cultures in 400 pregnant or puerperal patients Culture results (colonies/mi.) Gram smear

0

< 10.000

Negative Positive

253 0 253

56 7 63

Total

I

< 100.000

> 100.000

4

2

15

63 65

19

Total

3i5 85 400

996 Vicchi

first semiquantitative culture was negative despite symptoms of urinary tract infection, we obtained a second catheterization specimen one ~Neek later, in an attempt to detect infection resulting from one-time catheterization. In this study 44 women received no antibacterial prophylaxis; 69 received vesical instillation of 5 mi. 0.2 per cent nitrofurazone solution diluted in 50 mi. sterile distilled water; 20 received nitrofurazone solution as above, plus an urethral suppository of nitrofurazone; 20 received nitrofurantoin orally, 100 mg. three times a day for 3 days; 20 the same drug and dosage for 2 days; 20 sulfamethoxypyridazine by mouth, 1 Gm. on the first day followed by 0.5 Gm. daily for 3 days; and 20 received demethylchlortetracycline orally, 250 mg. three times a day for 3 days. Results. In the first group (30 women) correlation of culture counts of clean-voided midstream specimens with those of catheterization specimens obtained the following day showed less than 10,000 bacteria per milliliter in both specimens in 25 cases, and over 100,000 in both specimens in 2 cases. Results of the two specimen-collecting methods thus agreed in 27 (90 per cent) of the 30 cases. The catheter-collected specimen was negative (less than 10,000 organisms per milliliter of urine) while the voided specimen was positive in 3 cases; in 2 of these the voided specimen showed 100,000 and in the third 26,000 bacteria per miliiiiter. According to these findings we would expect an incider:ce of false positives of at least 10 per cent, v;hen specimens of urine are not obtained by catheterization. In the second study, in which repeat urine specimens were taken one week after collection of sterile specimens, both by catheter, at least 3 of the 44 patients in whom no prophylactic measure was used apparently contracted clinically significant infection: the second specimen showed over 100,000 bacteria per milliliter. Only one of the 169 patients in whom antibacterial prophylaxis was used was so infected. The difference in frequencies is statistically valid at the 95 to 98 per cent confidence level (chi square test) .

Df'cr-mhrl l. 1~Jb ~ •\m. 1- Obst, & Gynr<

This result supports the use of such treatment, as reported by ourselves 1 n and others.'' Although one of the 69 patients who received intravesical instillation of nitrofurazone showed significant bacteriuria one week later, we feel that the topical use of nitrofurazone is well suited for this purpose in obstetric patients; it is not systemic, and it is applied by the physician. All oral prophylactic medication seemed to be equally effective. Nitrofurantoin was associated with nausea and emesis in a small percentage of patients, while the two other drugs were well tolerated. Sulfamethoxypyridazine is the more economicaL Treatment of urinary tract infection

Methods. This study group consisted of 101 consecutive obstetric patients at the Hospital, treated during a period of 14 months for urinary tract infection of various degrees of severity and chronicity. Diagnoses were based on bacteriuria-pyuria, flank pain, dysuria, chills, and fever, as well as the results of comprehensive clinical tests, cystometry, Stemheimer-Malbin examination of urine sediment, and repeated cultures. Intravenous pyelography using diatrizoatc (Urografin) was performed in all puerperal patients with pyelonephritis. In each case a complete series of 3 films was taken, first a plain film, followed by one at 5 and one at 15 minutes after injection of the contrast agent. We performed urographic studies routinely within 4 weeks post partum (except in a pregnant patient suffering from colic due to urolithiasis) . In cystometric studies carried out in 25 of these patients the bladder was filled by rapid drip using tepid fluid, and vve recorded: ( 1) the desire to urinate, ( 2) the bladder capacity, and (3) the maximum pressure developed, in order to determine the predominating changes. We used this examination also to measure residual urine, if any. Pyelonephritis was diagnosed in 56 of our patients. Nineteen of these 56 were also pre-eclamptic, 9 were eclamptic, and 11 had other complications. Of the 45 patients without confirmed pyelonephritis 10 were pre-

Urinary tract infections

Volume 99 Number 7

997

Table II. Results of antibacterial treatment in 99 obstetric patients Treatment

Nitrofurantoin, oral Nitrdurantoin sodium (intramuscular) Nitrofurantoin sodium (intravenous) Demethyichlortetracycline and sulfamethoxypy.ridazine Chloramphenicol Total *Sev~n

Bacteriologic cure

9 9

Bacteriologic failure

2 1 8*

35 17

9t

4

5~

74

25

Total 11 10 43 26 9

99

of th€'se \'\!ere cured with chlorarnphenicol therapy.

tThree of these were cured with nitrofurantoin therapy.

tTwo of these were cured with nitrofurantoin therapy.

eclamptic and 12 suffered from other complicating conditions including urolithiasis, ureterocele, renal tuberculosis, endometriosis, appendicitis, cholelithiasis with cholecystitis, or pneumonia. Urine culture revealed Escherichia coli in 47, coliforms and other gram-negative organisms in 21, Staphylococcus aureus hemolyticus in 13, Proteus species in 9, and Aerobacter aerogenes in 8. Streptococcus hemolyticus, Mycobacterium tuberculosis, Diphtheroids, and Shigella species were noted in a few patients each. In vitro sensitivity tests were carried out routinely with all common antibiotics as well as several sulfonamides and ni.trofurans. Of the bacterial isolates in this series 75 per cent were found to be sensitive to nitrofurantoin, 43 per cent to chloramphenicol, 42 per cent to kanamycin, and 40 per cent to tetracyclines. Less than 20 per cent of the strains tested were sensitive to neomycin, penieillin, novobiocin, erythromycin, and other narrow spectrum antibiotics. Antibacterial treatment sometimes was initiated empirically before results of sensitivity testing were available. Therapy was continued usually for 5 days or more. We did not consider a patient to be cured bacteriologically until the colony count was repeatedly below 10,000 per milliliter. Results. Cystometric studies revealed increased bladder capacity and decreased detrusor tone (pressure) in 15 of the 25 patients in whom this procedure was carried out.

The bacteriologic results of treatment with the major drugs used in this study are shown in Table II for 99 patients who were available for complete follow-up study. Although 25 failures were observed, only 13 patients \Vere not cured by any or all of the drugs used; in the other 12 the infection could be cured by an alternative agent, as shown m the footnotes to Table II. Overall we observed bacteriologic cure m 75 per cent (74 of 99) of the patients. Nitrofurantoin efficacy (all routes combined) was somewhat greater than this: about 83 per cent (53 of 64 patients) . The other antiinfective agents used seemed Jess effective, although the number of observations was so small as to make this conclusion only tentative. Cases were classified as "Improved" when symptoms of infection clearly abated under therapy, but bacteriologic cure could not be confirmed by follow-up. Adopting "Satisfactory" as a heading to include both '"Cured" and "Improved" results, and combining the three routes of nitrofurantoin administration, the two most frequently used medications can be compared in a somewhat larger group of patients (Table III). Nitrofurantoin therapy resulted in a greater proportion of satisfactory results, and this difference reached the 95 per cent to 98 per cent confidence level (chi square test). Most of our patients were treated with nitrofurantoin or nitrofurantoin sodium because of the high initial in vitro sensitivity of common pathogens to this agent, and be-

998 Vicchi

Lh·n·mbet J.

Table III. Results of nitrofurantoin and demethylchlortetracyclinesulfamethoxypyridazine thera py-combinecl clinical and bacteriologic criteria ----------------

Result _ ,

j_ _

1

Satisfactory 1Failure I Therapy

Nitrofurantoin Demethy lchlortetracycline-sulfamethoxypyridazine Total

lY~t'

\nr. I. t )h;... t _ .\' Cpv·t

I

No. j %

(No.)

Total

65 18

87 69

10 8

75 26

83

82

18

101

cause better results had been obtained with this drug in practice, before the present study was undertaken, than with other antibacterials in the treatment of urinary tract in1') 13 . ., fection (the present author' s expenence . reports by others 8 ' 14 ' 15) . confi rmmg The principal disadvantage of nitrofurantoin oraiiy administered is nausea and vomiting in a small percentage of patients. The intravenous route (using the sodium salt) minimizes intolerance reactions while offering the advantage of rapid saturation of tissues with the drug. In a total of 75 patients treated with nitrofurantoin by all routes (because follow-up observations were not available, not all these patients are included in Table II), 4 experienced relatively severe ~rrastric intolerance. Treatment, . however. was successfully completed m all 4 case~. Two patients tolerated nitrofurantoin sodium well by the intravenous route after being nauseated by oral nitrofurantoin. One patient experienced considerable nausea and emesis throughout 5 days of intravenous therapy, but not so severely as to require interruption of treatment. The fourth patient was severely nauseated, with vomiting, after the first intravenous dose; because the micropathogen isolated from this patient showed high sensitivity to nitrofurantoin and resistance to aii other antibacterial agents, oral nitrofurantoin therapy was tried and successfully completed. Case report. L. S., aged 27, in the sixteenth week of pregnancy, suffered from hyperemesis

gravidarum; ""Vt're dysuna and In·qu(·ttn followed catheterization one month pn·viously. St·vere pyuria was found, with over 200,000 E. cnli per milliliter of urine. Intravenous py•·lography showed considt>rablt> left ureteral dilatation. The E. coli strain was very sensitive to nitrofurautoin or kana1nycin. Oral treatineni vvith nitr(•~ furantoin was discontinued after the first 2 doses because nausea and w;miting werP intensi-

fied. Intravenous nitrofurantoin sodium solution (360 mi. per day), however, was tolerated surprisingly well, with virtually complete dis~p­ pearance of signs and symptoms, and negauve urine cultures. The urine remained sterile at follow-up 5 weeks later, and at parturition, when the patient was deliven·d nf a normal infant.

In other women, hyperemesis gravidarurn was not necessarily aggravated by oral nitrofurantoin therapy. Four such patients tolerated the drug without gastrointestinal side effects; and the same was true of a patient who suffered from nausea and vomiting before treatment. due to cholelithiasis and cholecystitis. The intramuscular route is more convenient than the intravenous but may be more painful, and we found that adding procaine to the injection increased localized pain. Five of the 11 patients treated with nitrofurantoin sodium intramuscularly complained of severe pain at the injection site after one or more of the injections, and one of these also showed infiltration of the local tissues. Pain was not so severe as to interrupt treatment. The only other side effect seen with any .of the drugs used in this study \vas a relatively severe allergic skin reaction to sulfamethiazole, in one of the 7 patients treated with this agent. Effeds on the fetus. Neonates or infants were examined carefully for conditions possibly attributable to drug side effects. In the 58 cases in vvhich antibacte-rial agents were used during pregnancy, there was 1 case of spontaneous abortion (third month), 1 stillbirth, 1 premature birth, 1 icterus of unknown etiology, 1 congenital coxofemoral dislocation, and 4 instances of low birth weight. These frequencies are within the ranges usually seen at our institution. Spe-

Volume 99 Number 7

cific pediatric examination of 29 of the infants revealed no congenital metabolic errors

REFERENCES I. Kass, C. H.: Arch. Int. Med. 105: 194, 1960.

2. Rummel, H. H., and Walsch, E.: Ztschr. Geburtsh. u. Gynak. 160: 105, 1963. 3. Forkman, A.: Acta obst. et gynec. scandinav. 4:3: 35, 1964. 4. Low, J. A., Johnston, E. E., McBride, R. L., and Tuffnell, P. G.: AM. J. 0BsT. & GYNEC. 90: 897, 1964. 5. Kass, C. H.: Ann. Int. Med. 56: 64, 1962. 6. LeBlanc, A. L., and McGanity, W. J.: Texas State J. Med. 60: 137, 1964. 7. Kincaid-Smith, P., Bullen, M., Mills, ]., Fussdl, U., Huston, N., and Goon, F.: Lancet 2: 61, 196+.

Urinary tract infections

999

or internal malformations other than as just stated.

8. Semmens, J. P.: Consultant 4: 30, 1964. 9. Sternheimer, R., and Malbin, B.: Am. J. Med. 11: 312, 1951. 10. Vicchi, M. F.: Proc. VIII Congr. Panamer. Ural., Bogota, 1963. II. Levin, J.: Ann. Int. Med. 60: 914, 1964. 12. Vicchi, M. F., and Abdo, G. T.: Rev. Soc. Ecuador. Gin. Obst. 5: 1961. 13. Vicchi, M. F., Jijon, A. M., Espinosa, F., Abdo, G. T., and Larrea, A.: Proc. Congr. Ecuador. ginec. e obst., 1963. 14. Catlow, C. E.: J. Ural. 86: 3.'i1, 1961. 15. Sanjurjo, L. A.: Tr. Am. A. Genitour. Surg. 54: 73, 1962.