Volume 155 Number 3

The association of brown amniotic fluid at the time of second-trimester genetic amniocentesis with a history of vaginal bleeding is well known. Legge l reported that the pigments contributing to the discoloration include hemoglobin A, hemoglobin F, and myoglobin with only myoglobin being of poor prognostic significance.

It is surprising that purely maternal bleeding oc­curred into the amniotic sac. This phenomenon can occur if maternal blood dissects into the potential space between the chorion and decidua and then bursts through the amniotic membrane and enters the am­niotic cavity. The site of entry may then seal over.

As far as we are aware, this is the most advanced example of an angular pregnancy examined sono-

Maternal liemorrhage into amniotic sac

graphically. It is disappointing that the appearances were identical tb those of a normal intrauterine preg­nancy. Although clinically the pregnancy was felt to be asymmetric in location, it appeared centrally placed on the sonogram. Interstitial pregnancy is acknowledged to be a difficult sonographic diagnosis. 2

REFERENCES 1. Legge M. Dark brown amniotic fluid-identification of

contributing pigments. Br J Obstet Gynaecol 1981 ;88: 632-4.

2. Graham M, Cooperberg PL. Ultrasound diagnosis of in­terstitial pregnancy: findings and pitfalls. JCU 1979;7:433.

Therapy of candidal vaginitis: The effect of eliminating intestinal Candida

Nystatin Multicenter Study Group

A total of 258 patients with candidal vulvovaginitis, all of whom also exhibited Candida organisms in the

rectum, were treated for 1 week with vaginal tablets only (nystatin or clotrimazole) or with both vaginal and

oral tablets (nystatin). Mycologic and symptomatic responses were superior for the group receiving combined intravaginal-oral therapy; the vaginas of 88% of those treated by both routes were cleared of

Candida, as compared with 75% of those receiving only intravaginal medication (p < 0.05). Nystatin and

clotrimazole were equally effective. When the 258 patients, regardless of treatment regimen, were grouped into those whose intestinal tracts after therapy contained Candida or those free of Candida, the response

rates of the vaginal infection at all follow-up examinations favored the latter group (p < 0.05 to < 0.001). Vaginal infection recurred in 19.7% of patients treated only intravaginally and 14.7% of those receiving combined therapy. These results suggest the value of eliminating any intestinal reservoir of Candida when

treating patients with candidal vulvovaginitis. (AM J OSSTET GVNECOL 1986;155:651-5.)

Key words: Candida albicans, intestinal candidiasis, vaginitis, nystatin, clotrimazole

Many factors, including use of oral contraceptives, pregnancy, and treatment with broad-spectrum anti­biotics or corticosteroids, predispose the patient to vul­vovaginal candidosis and complicate its management. It long has been suspected that the gastrointestinal tract also may playa role by affording a reservoir of Candida

organisms capable of initiating or perpetuating this common and distressing infection. 1-4 However, the few studies conducted to date have failed to demonstrate

Individual members of the Nystatin Multicenter Study Group are listed at the end of the article.

Received for publication February 6, 1986; accepted April 22, 1986. Reprint requests: A. von Gueltlingen, M.D., Richard Strauss Strasse

45, 8000, Munich 80, Germany.

that the vaginitis is better managed by suppressing this intestinal reservoir with an antifungal agent. 5

.6

The present multicenter study was undertaken to explore further the possibility that patients with can­didal vaginitis will benefit if local (intravaginal) therapy is supplemented by removal of C. albicans from the lower intestinal tract. Ancillary to this primary objec­tive, we sought to determine whether intravaginally ad­ministered clotrimazole has any significant advantage over the long-established use of nystatin, as has been reported by SOme-IO but not by others. ll-I3

Material and methods

Case material was provided by 10 gynecologists lo­cated in six countries. The clinical diagnosis of candidal vaginitis was confirmed in all 271 patients by a positive

651

652 Nystatin Multicenter Study Group

Table I. Composition of study population

Group J-

September 1986 Am J Obstet Gynecol

oral and intravaginal nystatin Group 2-

intravaginal nystatin Group 3-

intravaginal clotrimazole

No. of patients Mean age (yr)

Range Married Single Divorced or widowed Pregnant Oral contraceptive use No. of candidal vaginitis episodes

in previous 12 months I 2-4 >4 Unknown

83 30

17-92 63 16 4

II 23

7 67

4 5

86 31

16-62 63 21

2 II 25

6 66

8 6

89 30

17-52 62 24

3 15 23

4 69

6 10

Table II. Response of candidal vaginitis to intravaginal therapy

Visit 2

% n

Clinical response* Clotrimazole (n = 89) 42 35/83 Nystatin (n = 86) 32 27/84

NS Negative vaginal mycology

Clotrimazole (n = 89) 61 51183 Nystatin (n = 86) 69 58/84

NS

NS, Not statistically significant.

*Disappe,!rance of signs and symptoms of infection.

potassium hydroxide examination of a vaginal swab specimen and also in all but 11 by a positive culture of C. albicans. Wet mounts and hanging drop preparations were obtained to exclude patients with vaginal infec­tions caused by Gardnerella vaginalis and Trichomonas vaginalis, respectively. For the purpose of this study, material from a pretreatment rectal swab from each patient had to exhibit the pseudohyphae and budding yeasts of Candida on potassium hydroxide examination or produce a positive culture.

All patients qualifying for admittance to the trial re­ceived combined local (i·ntravaginal) and oral treatment daily for 1 week as follows: Oral: group 1, nystatin; group 2, placebo; group 3, placebo. Local: group 1, nystatin; group 2, nystatin; group 3, clotrimazole.

Participating physicians assigned patients to group 1,2, or 3 according to a ra)1dom design. The oral med­ication (500,000 U nystatin tablets or matching placebo tablets) was given as two tablets three times daily, that is, 3,000,000 U daily. A single vaginal tablet containing either 100,000 U of nystatin or 100 mg of clotrimazole was inserted each night at bedtime. The I-week course of treatment was not interrupted for menses.

Visit 3 Visit 4

% n % n

57 46/80 72 55176 53 41178 68 50174 NS NS

74 59/80 75 57176 76 59178 76 56174 NS NS

Follow-up clinical and mycologic examinations (po­tassium hydroxide preparation and culture) to assess the response of the vaginitis were accomplished 1 to 7, 8 to 21, and 22 to 49 days after completion of therapy. At each of these time points (hereafter referred to as visits 2, 3, and 4, respectively) material from a rectal swab also was examined mycologically (potassium hy­droxide preparation and culture).

Statistical analyses of the therapeutic responses ob­served with the three treatment regimens were per­formed with use of the Mantel-Haenszel approach to the analysis of rates and proportions. 11 Fisher's exact test was used to analyze all 2 x 2 tables.

Results

Of the 271 patients with proved vaginal candidiasis and coexisting intestinal Candida, 258 satisfactorily completed the study (83, group 1; 86, group 2; 89, group 3). Five patients in group 1 were lost to follow­up, as were one each from groups 2 and 3. Three patients each were excluded from groups 1 and 3 for violations of the study protocol. Most of the 258 pa­tients had experienced at least two prior episodes of

Volume 155 Number 3

Intestinal Candida and candidal vaginitis 653

Table III. Clinical and mycologic responses in candidal vaginitis

Intmvaginal plus oml Intravaginal thempy* therapyt

% I n % I n Significance

Clinical response Visit 2 37 621167 48 37177 NS Visit 3 55 871158 80 61176 P < 0.001 Visit 4 70 1051150 87 63172 p< 0.01

Negative vaginal mycology Visit 2 65 1091167 79 61177 P < 0.05 Visit 3 75 1181158 88 67176 P < 0.05 Visit 4 75 1131150 86 62172 NS

*Groups 2 and 3 combined. tGroup I.

Table IV. Influence of continuing candidal colonization of intestinal tract* on response to therapyt for candidal vaginitis

Visit 2 Visit 3 Visit 4

% n % n % n

Clinical response Colonization 25 301118 37 37/99 52 46/88 No colonization 55 691126 82 1111135 91 1221134

P < 0.001 P < 0.001 P < 0.001 Negative vaginal mycology

Colonization 52 611118 69 68/99 65 57/88 No colonization 86 1091126 87 117/135 88 1181134

P < 0.001 P < 0.05 P < 0.001

*Determined at each visit by potassium hydroxide examination and culture of rectal swabbings. tClotrimazole vaginal tablets; nystatin vaginal tab:ets with or without nystatin oral tablets.

candidal vaginitis within the past year. Demographic

and other characteristics of the three treatment groups

generally were comparable (Table I).

The clinical and mycologic responses to therapy with only local nystatin (group 2) or c1otrimazole (group 3)

were similar at all three follow-up examinations (Table

11*); therefore the findings for these two groups, that

is, local therapy only, have been combined for com­

parison with group I (local therapy plus oral nystatin). Table III compares the clinical and mycologic re­

sponses to intravaginal plus oral nystatin (group I) and

intravaginal therapy only (groups 2 and 3 combined). Although the differences were not statistically signifi­

cant at every visit, group I consistently had a better response rate.

Rectal swab specimens obtained after therapy were negative for Candida in more than 70% of patients in

group I; the rectal contents also became free of Candida (presumably spontaneously) by the first follow-up ex­amination in 42% of the combined group 2 and 3 pa-

*The numbers of patients at visits 2, 3, and 4, as shown in Tables II to IV, differ somewhat because not all patients pre­sented for all follow-up visits.

tients. Therefore we divided the total study population into those whose intestinal tracts continued to show this

organism and those cleared of it. As shown in Table IV, at all three follow-up examinations both the my­cologic and clinical response rates of the vaginitis were

significantly higher when the intestinal tract was free

of Candida.

Mycologically confirmed recurrence of candidal vul­

vovaginitis occurred by the second or third follow-up visit in II %, 22%, and 16% of group I, 2, and 3 pa­tients, respectively, who were originally "cured" (that

is, had a negative vaginal mycologic examination at the first follow-up visit). Recurrences were observed in

14.7% (16/109) of all such patients whose rectal potas­sium hydroxide preparations and cultures at the first follow-up were negative and in 19.7% (12/61) of those whose rectums contained Candida (p > 0.05).

All three drug regimens were well tolerated.

Comment In the present study monotherapy with nystatin or

c1otrimazole vaginal tablets produced similar responses

in patients with candida I vulvovaginitis. This was so

despite the fact that nystatin therapy was continued for

654 Nystatin Multicenter Study Group

o Candida Present in Intestinal Tract

100

~ 0 ...... >-

" 75 9 0 () >-~ ...J <{ 50 z a ~ w > ~ 25

" w z

C1J No Candida in Intestinal Tract

p<O.001

September 1986 Am J Obstet Gynecol

N= 96 71 22 55 88 70 11 65 80 70 8 IVg IVg + PO IVg IVg + PO IVg IVg + PO

~VISIT2--1 ~ VISIT 3-----1 ~VISIT4~

Fig. 1. Response to therapy of candidal vaginitis as related to presence of C. albicans in the intestinal tract. IVg = Intravaginal therapy only; IVg + PO = intravaginal plus oral therapy; open bar =

Candida present in intestinal tract; hatched bar = no Candida in intestinal tract.

only 1 week, a short course that in previous studies provided responses inferior to those obtained with clo­trimazole. 7

.10 Because of the similarity of the findings

with both drugs the two treatment groups could be merged for comparison with the treatment regimen combining intravaginally and orally administered ny­statin. Predictably the latter regimen cleared the intes­tinal tract of the Candida organism in almost 70% of patients (negative rectal culture and potassium hy­droxide preparation at both first and second follow-up visits.) Close to 35% of the patients receiving only in­travaginal therapy also exhibited Candida-free rectal specimens at both follow-up visits, an unexpected find­ing also reported earlier by Lebherz et al. l ; among patients receiving local therapy for candidal vulvo­vaginitis.

The response rate of the patient population that re­ceived the combined local-oral therapy consistently was superior to that of those receiving only local therapy. Among the patients in both treatment groups the clin­ical and mycologic response rates were similarly and significantly improved when the intestinal tract was free of Candida (Fig. 1). In fact, the odds l6 of obtaining a mycologically confirmed cure favored patients whose rectums had been cleared of this organism by at least two to one. The fact that this situation prevailed twice as frequently in the combined local-oral treatment group accounts for the better results obtained with the

latter therapy. The better response rate in patients whose intestines became Candida-free is consistent with the earlier observation of Lambotte et al. 17 that patients with candidal vaginitis responded better to local ther­apy if the organism was not present in their rectums before therapy.

It also was of interest to compare the rapidity of response of the vaginitis to therapy and the frequency of recurrence in patients whose intestinal tracts did or did not afford a reservoir for the Candida organism. To this end we reviewed the findings at the first follow­up visit (l to 7 days after completion of therapy) for all patients who ultimately responded to therapy, that is, vagina was negative for Candida at the second follow­up (8 to 21 days after therapy). Clinical and/or myco­logic responses were observed in 74% of the 88 patients exhibiting Candida in rectal specimens at either the sec­ond or third follow-up examination and in 94% of the 85 patients with negative rectal specimens at both visits (p < 0.01). Thus it appears that not only is the response of the vaginitis to therapy more consistent, but also it occurs more rapidly when there is no so-called intestinal reservoir of Candida.

Recurrences of the vaginitis were observed in only 28 of the 170 patients whose vaginas were negative for Candida at the end of therapy. The briefness of the follow-up period (up to 49 days) may account for this low recurrence rate, and in turn the latter may account

Volume 155 Number 3

for the lack of difference in the frequency of recur­

rences in patients whose rectums were colonized with

Candida and those who were free of this organism.

The present study, in which significantly better ther­

apeutic responses were obtained in vulvovaginal can­

didiasis when the intestinal tract was rid of C. albicans,

gives clinical significance to earlier observations of a

positive association between the presence of Candida in

the vagina and in the rectum. I. 2. 1. 18

Individual members of the Nystatin Multicenter Study Group are as follows: A. von Gueltlingen, M.D. (Munich, Germany), A. Ayala, M.D. (Mexico City, Mex­ico), R. A. Apelo, M.D. (Manila, Philippines), R. M. Ramos, M.D. (Manila, Philippines), F. Revollo, M.D. (Bogota, Colombia), M. Alfaro, M.D. (Guatemala City, Guatemala), M. Bueno, M.D. (Cali, Colombia), J. A. Erlingson, M.D. (Stockholm, Sweden), B. Friberg, M.D. (Halmstad, Sweden), L. G. Carlborg, M.D. (Halmstad, Sweden).

REFERENCES

I. Miles MR, Olsen L, Rogers A. Recurrent vaginal candi­diasis. JAMA 1977;238: 1836-7.

2. Hilton AL, Warnock DW. Vaginal candidiasis and the role of the digestive tract as a source of infection. Br J Obstet Gynaecol 1975;82:922-6.

3. Rohatiner J.J. Relationship of Candida albicam in the gen­ital and anorectal tracts. Br J Vener Dis 1966;42: 197-200.

4. de Sousa HM, Uden N van. The mode of infection and reinfection in yeast vulvovaginitis. AM J OBSTET GYNECOL 1960;80: 1096-1100.

5. Velupillai S, Thin RN. Treatment of vulvovaginal yeast infection with nystatin. Practitioner 1977;219:897-90 l.

6. Milne JD, Warnock DW. Effect of simultaneous oral and

Intestinal Candida and candidal vaginitis

vaginal treatment on the rate of cure and relapse in vag­inal candidosis. Br J Vener Dis 1979;55:362-5.

7. Tan CG, Milne LJR, Good CS, LoudonJDO. A compar­ative trial of six day therapy with clotrimazole and nystatin in pregnant patients with vaginal candidiasis. Postgrad MedJ 1974;50(suppll):102-5.

8. Higton BK. A trial of clotrimazole and nystatin in vaginal candidiasis. Postgrad Med J 1974;50(suppl 1):95-7.

9. Masterton G, HendersonJ, Napier IR, Moffet M. Six-day clotrimazole therapy in vaginal candidosis. Curr Med Res Opin 1975;3:83-8.

10. Marks H.J. A double-blind comparison of clotrimazole and nystatin vaginal tablets in candida vaginitis. Postgrad Med J 1974;50(suppl I): 105-8.

II. Eliot BW, Howat RCL, Mack AE. A comparison between the effects of nystatin, clotrimazole and miconazole on vaginal candidiasis. Br J Obstet Gynaecol 1979;86:572-7.

12. Svendsen E, Lie S, Gunderson TH, Lyngstad-Vik I, Sku­land .J. Comparative evaluation of miconazole, clotrima­zole and nystatin in the treatment of candidal vulvo-vag­initis. Curr Ther Res 1978;23:666-72.

13. Cartwright RY. Clotrimazole in the treatment of acute and 'resistant' vaginal candidiasis. Postgrad Med J 1974;50 (suppl 1):90-2, 1974.

14. Fleiss JL. Statistical methods for rates and proportions. New York: John Wiley & Sons, 1973.

15. Lebherz TB, Goldman L, Wiesmeier E, Mason D, Ford Le. A comparison of the efficacy of two vaginal creams for vulvovaginal candidiasis, and correlations with the presence of Candida species in the perianal area and oral contraceptive use. Clin Ther 1983;5:409-16.

16. Bishop YMM, Fienberg SE, Holland PW. Discrete mul­tivariate analysis theory and practice. Cambridge, Mas­sachusetts: MIT Press, 1975.

17. Lambotte R, Dogniez B, Sandront-Degee M. Influence of rectal infections caused by Candida albicans on the thera­peutic results obtained in vaginal candidosis. Mykosen 1978;21(suppl 1):311-3.

18. Anyon CP, Desmond FB, Eastcott DF. A study of Candida in one thousand and seven women. NZ MedJ 1971;73:9-13.

Treatment of cornual pregnancy with methotrexate: Case report

Mary C. Brandes, M.D., David D. Youngs, M.D., Donald P. Goldstein, M.D., and Tim H. Parmley, M.D.

Portland, Maine, and Baltimore, Maryland

Presented is the case of a woman with a cornual pregnancy documented by ultrasound and laparoscopy

and successfully treated with an 8-day course of methotrexate. (AM J OBSTET GVNECOL 1986;155:655-7.)

Key words: Ectopic pregnancy, methotrexate

From the Department of Obstetrics and Gynecology, Maine Medical Center, Harvard Medical School and the New England Tropho­blastic Disease Center, and the Division of Gynecologic Pathology, Department of Obstetrics and Gynecology, The Johns Hopkins M ed­ical Institution.

Received for publication March 11, 1986; accepted April 3, 1986. Reprint requests: David D. Youngs, M.D., Department of Obstetrics

and Gynecology, Maine Medical Center, Portland, ME 04102.

Successful treatment of ectopic pregnancy by che­

motherapeutic means has been infrequently recounted

in the literature. However, several authors have sug­

gested its usefulness in appropriate cases where surgical

managment might impair future infertility I. 2 or be of

significantly greater risk to the patient. 2 Presented is a

655