Genitourin Med: first published as 10.1136/sti.62.4.235 on 1 August 1986. Downloaded from

Genitourin Med 1986:62:235-9

Treating pelvic inflammatory disease with doxycycline and or and metronidazole.

P K HEINONEN,* K TEISALA,* R PUNNONEN,* R AINE,t M LEHTINEN,i: A MIETT1NEN,§ AND J PAAVONEN¶

From theDepartments of*Obstetrics and Gynaecology, t Pathology, and t Medical Microbiology, University Central Hos,pltal, and the Departments of §Biomedical and I Clinical Sciences, University of Tampere, Tampere, Finland

SUMMARY The best way oftreating pelvic inflammatory disease (PID) is not known. The clinical response to two treatment regimens (penicillin plus metronidazole v doxycycline plus metronidazole) was studied in 33 patients with PID confirmed by laparoscopy and endometrial biopsy. The overall failure rate, according to the criteria used in this study was five of 11 (45%) women with chlamydial PID, none of six women with gonococcal PID, all of four women with chlamydial gonococcal PID, and three (25%) of 12 women with non-chlamydial non-gonococcal PID. The failure rate with penicillin plus metronidazole treatment was unacceptably high (53%), and significantly higher than that with doxycycline plus metronidazole ( 1 9%) (p=0'038). In most cases the microbiological and histopathological evaluations identified a probable explanation for the poor response to the treatment regimen used. http://sti.bmj.com/

Introduction laparoscopically diagnosed salpingitis or histopatho- logically diagnosed , or both. The Acute pelvic inflammatory disease (PID) is usually an presence of salpingitis was assessed using the lapar- ascending polymicrobial ofthe endometrium scopic criteria ofmild, moderate, or severe salpingitis,8 and fallopian tubes.'-3 Though laparoscopy and and a single strip endometrial biopsy specimen was endometrial sampling have improved its diagnosis,45 obtained, as described.5 The histopathological on September 27, 2021 by guest. Protected copyright. the microbial aetiology of tubal infection is rarely diagnosis ofendometritis was based on the presence of known. Thus antimicrobial treatment should cover all plasma cells on endometrial biopsy. Endometritis was common causative organisms.67 Previous studies have classified as mild, moderate, or severe.5 Cervical, shown that roughly 15%-20% of patients with acute endometrial, tubal and peritoneal cultures for isolation PID fail to respond to initial treatment, and in up to of trachomatis, Neisseria gonorrhoeae, 30% the disease recurs.2 , Ureaplasma urealyticum, This study aimed to evaluate the efficacy of two herpes simplex virus, and anaerobic and facultative antimicrobial combinations commonly used in the were obtained as described previously.4 We treatment of acute PID. classified all cases according to the culture results for C trachomatis and N gonorrhoeae as follows: Patients and methods chlamydial PID, if C trachomatis but not The study group comprised 33 women who had Ngonorrhoeae was isolated from any site: gonococcal PID, if N gonorrhoeae but not C trachomatis was isolated from any site: chlamydial gonococcal PID, if Address for reprints: Dr P K Heinonen, Department ofObstetrics and both organisms were isolated; and Gynaecology, University Central Hospital, SF-33520 Tampere, non-chlamydial Finland non-gonococcal PID, if neither organism was isolated. Accepted for publication 20 December 1985 Antimicrobial treatment was started intravenously 235 Genitourin Med: first published as 10.1136/sti.62.4.235 on 1 August 1986. Downloaded from

236 P K Heinonen, K Teisala, R Punnonen, R Aine, M Lehtinen, A Miettinen, and J Paavonen immediately after laparoscopy. The patients were TABLE I Selected characteristics and clinical findings in unselectively treated either with the combination of women treatedfor acutepelvic inflammatory disease (PID) penicillin plus metronidazole or doxycycline plus metronidazole. The patients with an even year ofbirth Penicillin Doxycycline were given intravenous benzylpenicillin 1 8 g three and and times a day plus metronidazole 500 mg three times a Observation metronidazole metronidazole day for 48 hours, followed by 660 mg phenoxymethyl penicillin plus 400 mg metronidazole three times a day No of patients 17 16 Age (years)* 26 (9) 23 (4) by mouth to complete 14 days of treatment. The Nulliparous 8 8 women with an odd year of birth were given Contraception intravenous doxycycline hydrochloride 100 mg twice IUCD 9 5 daily plus metronidazole 500 mg three times a day Hormonal 2 4 Other methods 4 4 intravenously for 48 hours, followed by 150 mg None 2 3 doxycycline once daily plus metronidazole 400 mg History of PID 3 7 three times per day by mouth to complete 14 days Duration of symptoms (days)* 11 (7) 8 (8) of treatment. Clinical severity score* 11 (5) 11 (5) Erythrocyte sedimentation 34 (23) 35 (24) A clinical severity score was derived by grading the rate (mmlhour)* presence of cervical motion and uterine and adnexal C reactive protein (mg/I)* 67 (54) 53 (44) tenderness on a scale of 0 (absent to normal) to 3 (severe), as described previously.9 An intrauterine * Values are mean (SD). contraceptive device (IUCD) was present in 14 patients, which was removed before the start of treat- ment. The clinical examination was repeated 48-72 TABLE II Occurrence of treatment failures in relation to hours, seven days, and 14 to 16 days after treatment isolation of or Neisseria had been started. Cervical cultures for C trachomatis gonorrhoeae in 33 patients treated for acute pelvic and Ngonorrhoeae were repeated at the 14 to 16 days inflammatory disease (PID) visit. Erythrocyte sedimentation rate, serum C reactive protein concentration,'0 and white cell count were determined initially. The erythrocyte sedimenta- No offailures/No treated with: and white cell count were after 14 tion rate repeated Penicillin Doxycycline days of treatment. and and Treatment failure was defined as follows: no ClassUication ofPID metronidazole metronidazole improvement in the clinical severity score during http://sti.bmj.com/ pelvic examination performed 48-72 hours after the Chlamydial 3/4 2/7 start of treatment; presence of pelvic mass, clinical Gonococcal 0/5 0/1 > Chlamydial gonococcal 4/4 0/0 severity score 5, positive cervical culture for Non-chlamydial non-gonococcal 2/4 1/8 N gonorrhoeae or C trachomatis, or erythrocyte Total (%) 9/17 (53) 3/16 (19) sedimentation rate higher than halfofthe initial level at the 14 to 16 days examination. ^ Fisher's exact test was used for statistical on September 27, 2021 by guest. Protected copyright. comparisons. Table III shows the microbiological findings and Results severity of endometritis and salpingitis in patients treated with penicillin plus metronidazole. Of eight Table I shows that the demographic and clinical women who had PID associated with C trachomatis characteristics ofthe patient groups were comparable. infection, seven (88%) failed to respond to penicillin Seventeen patients were treated with penicillin plus plus metronidazole treatment (tables II and III). In metronidazole, and 16 were treated with doxycycline both patients with non-chlamydial non-gonococcal plus metronidazole. PID in whom treatment failed , or Table II summarises the response to treatment in 17 influenzae, was isolated from the fallo- patients who received penicillin plus metronidazole. pian tubes (table III). In one case H influenzae was According to our criteria the treatment had failed in isolated from tubal aspirate, and in the otherE coli was nine (53%) cases, including three (75%) of four isolated from the tube and the endometrium. In one patients with chlamydial PID, four (100%) women patient (case 7) who responded clinically E coli was with chlamydial gonococcal PID, and two (50%) of isolated from the fallopian tube. Anaerobic bacteria four women with non-chlamydial non-gonococcal were isolated from the endometrium in seven cases. PID. Patients who had not responded to penicillin plus Genitourin Med: first published as 10.1136/sti.62.4.235 on 1 August 1986. Downloaded from

Treating pelvic inflammatory disease with doxycycline orpenicillin (both with metronidazole) 237

3 St @ ~~oo o 2 o o o o o o o o o I o o o o o o o o I o

~~~ ~~~~~~~~~~0- 4

4)~~~~ ~ ~ ~ ~ ~ ~ ~ ~~~~()4

-4 .~y~.iBa; ; e; ; WXZ X

41)00 2 CO 2 22 Z 4 Z 4.o) Z)

00 c) 0 0 Z .4. I oI*... > o > > > > > ~~~~~~~~ ~ ~ ~ ~ 6 .4 |.4 44 40. 4)04 4))44 44

S:~~~~~~z z zz

w .00 0 o

0~~~~~~~~~~~~~~~~~

130~~~~~~~~~~~~~~~~~~

0~~0 http://sti.bmj.com/ 4) 1 1 - 0 0 o

u '4 u 11 ;h g2U t040~ ~ ~ ~~

i C4 >~~~~~~~44 t%. t-£0Ob_^UUbo U.-; _X Q

52_Y-£OO_uu~0 on September 27, 2021 by guest. Protected copyright.

cn 0~~~~~~~~~~~

0qM'dn~ -WO C4) U ---q-114--4-- - - -C4 . . eq eq .. . e en e en 0' Genitourin Med: first published as 10.1136/sti.62.4.235 on 1 August 1986. Downloaded from

238 PKHeinonen, K Teisala, R Punnonen, R Aine, MLehtinen, A Miettinen, andJPaavonen metronidazole were subsequently treated with penicillin plus metronidazole (5 3%). Our results agree doxycycline (three cases), or with the combination of with those ofprevious studies.15-20 In two ofthree cases sulphadiazine plus trimethoprim (four cases), or with with no response to doxycycline plus metronidazole amoxycillin (two cases). C trachomatis was isolated from the cervix, Table III summarises the findings and clinical endometrium, and fallopian tubes. Severe plasma cell response in the 16 patients who were treated with endometritis was detected by endometrial biopsy. The doxycycline plus metronidazole. This treatment dose, or the duration of the treatment, may have been regimen failed in three ( 19%) patients, including two insufficient. Resistance to doxycycline among of seven (29%) with chlamydial PID and one of eight C trachomatis strains has not been described. (12%) patients with non-chlamydial non-gonococcal Histopathological examination showed severe PID (tables II and III). Bacteroides melaninogenicus was isolated from the fallopian tubes in the patient with endometritis in 12 patients, and in eight of these non-chlamydial non-gonococcal PID who failed to C trachomatis was also isolated from the respond. She developed pelvic peritonitis with a tubal endometrium. Remarkably, severe endometritis was which was three times. detected in 50% of the patients who failed to respond , aspirated This deserves Actinomyces israelii was isolated from the removed to the antimicrobial treatment. finding IUCD. further study. Perhaps the presence of severe The severity ofendometritis seemed to predict treat- endometritis predicts the failure of conventional treat- ment failure slightly better than the severity of ment. In such cases, therefore, the treatment should be salpingitis. Thus treatment failure occurred in six of 12 more effective with higher doses and longer duration. Serial determinations of serum women (50%) with severe and three of C reactive protein endometritis, might also be helpful in monitoring response to the seven women (43%) with severe salpingitis (table III). treatment. As previous studies have shown a high rate of anaerobic bacteria in the peritoneal fluid, or endo- Discussion metrial cavity, of women with PID, we combined metronidazole with both the antimicrobial agents used Optimal treatment for PID is difficult to determine for to gain better coverage against anaerobic bacteria. In the following reasons: the microbial aetiology of tubal such studies, however, the specimens were obtained by infection is rarely known; a uniform objective clinical culdocentesis, or transcervically, so that vaginal scoring system to evaluate the short term response to contamination of the specimens might have treatment has not been widely used; and long term occurred.3 15 21 In this study we found a much lower follow up studies for tubal infertility after antimicrobial prevalence of anaerobic bacteria in the endometrial http://sti.bmj.com/ treatment are extremely difficult. In this study we used and tubal specimens. The use of metronidazole is the clinical scoring system for tendemess developed by probably unnecessary in the treatment of uncom- McCormack et al 9 to assess the clinical response to plicated PID without pelvic . On the other antimicrobial treatment. We also used laparoscopy, hand, in this study mixed facultative and anaerobic endometrial biopsy, and microbiological sampling of bacteria were commonly found in the endometrial the cervix, endometrium, tubes, and peritoneal cavity specimens, supporting the argument for using to obtain objective evidence of the presence, severity, metronidazole in the combination treatment, though on September 27, 2021 by guest. Protected copyright. and aetiology of PID. the possibility of cervicovaginal contamination is Both single drug and combination drug treatments extremely difficult to rule out when any transcervical have been used in the treatment of acute PID. Failure sampling method is used. rates of previous treatment trials have ranged from E coli, H influenzae, and B melaninogenicus were 0%-55% (mean 12%)." 12 The failure rate for the examrples of non-gonococcal and non-chlamydial combination of penicillin plus metronidazole in this organisms isolated directly from the fallopian tubes. study was unacceptably high (53%). This is undoub- Three such cases did not respond to the initial treat- tedly due to the fact that penicillin or metronidazole ment. These findings stress the importance of obtain- are not effective against C trachomatis. Our findings ing microbiological specimens directly from the agree with those of previous studies on women with inflamed tubes. Laparoscopy is necessary to obtain cervicitis.13 these specimens and to evaluate the severity of the Doxycycline plus metronidazole has been recom- disease. In two cases H influenzae was isolated from mended for the treatment of acute PID by the Centers tubo-ovarian abscesses when these abscesses were for Disease Control, United States ofAmerica.'4 This aspirated under laparoscopic control. In addition to combination is effective against both anaerobes and appropriate antimicrobial treatment and laparoscopic C trachomatis. In this trial the failure rate was 19%, damage of abscess, further surgical interventions werie which is significantly (p=0 038) lower than that for not required in any of the cases. Genitourin Med: first published as 10.1136/sti.62.4.235 on 1 August 1986. Downloaded from

Treating pelvic inflammatory disease with doxycycline or penicillin both with metronidazole 239 In conclusion, our study clearly shows that and nongonococcal infection and evaluation oftheir response to penicillin plus metronidazole is an inadequate treat- treatment with aqueous procaine penicillin G and ment for acute PID. Doxycycline plus hydrochloride. Sex Transm Dis 1977;4:125-31. mentPID. metronidazole ~~~~~~~~~~~~~10.Harmoinen A, Perko M, Gronroos P. Rapid quantitative deter- treatment showed a significantly lower ( 19%) clinical mination of C-reactive protein using LKB 8600 reaction rate failure rate. We were able to identify a reasonable analyzer. Clin Chim Acta 1981;111:1 17-20. microbiological or histopathological explanation for 11. Rees E. The treatment of pelvic inflammatory disease. Am J thetpoorpresponseroinimostomcases of treatment failure. 1 Obstet Gynecol 1980;138:1042-7. This emphasises the importance of extensively 12- critiqueBrunhamofRC.recentTherapytreatmentfor acutetrials.pelvicAminflammatoryJ Obstetdisease:Gynecola evaluating patients with suspected PID. 1984;148:235-40. 13. Stamm WE, Guinan ME, Johnson C, Starcher T, Holmes KK, This study was supported by grant number 7939/304/83 McCormack WM. Effect of treatment regimens for Neisseria from the Academy of Finland. gonorrhoeae on simultaneous infection with trachomatis. N Engl J Med 1984;310:545-9. Chlamydia 14. Centers for Disease Control. Sexually transmitted diseases References treatment guidelines 1982. MMWR 1982;31:43-62. 15. Monif GRG. Significance of polymicrobial bacterial superin- 1. Jacobson L, Westrom L. Objectivized diagnosis of acute pelvic fection in the therapy of gonococcal endometritis-salpingitis- inflammatory disease. Diagnostic and prognostic value of peritonitis. Obstet Gynecol 1980;55(suppl 5):154-161. routine laparoscopy. Am J Obstet Gynecol 1969;105:1088-98. 16. Andersson PO, Hackl H, Jensen P, Larsen KR. A comparison 2. Westrom L. Incidence, prevalence, and trends of acute pelvic of two different dosages of pivampicillin and doxycycline in inflammatory disease and its consequences in industrialized patients with gynaecological . Curr Med Res Opin countries. Am J Obstet Gyniecol 1980;138:880-92. 1980;6:513-7. 3. Sweet RL. Diagnosis and treatment of pelvic inflammatory 17. Gjonnaess H, Dalaker K, Urnes A, et al. Treatment of pelvic disease in the emergency room. Sex Transm Dis 1981 ;8: 156-63. inflammatory disease effects of and clindamycine. 4. Heinonen PK, Teisala K, Punnonen R, Miettinen A, Lehtinen Current Therapeutic Research 1981 ;29:885-92. M, Paavonen J. Anatomic sites of upper genital tract infection. 18. MonifGRG, Welkos SL, Baer H. Clinical response ofpatients Obstet Gynecol 1985;66:384-90. with gonococcal endocervicitis and endometritis-salpingitis- 5. Paavonen J, Aine R, Teisala K, et al. Chlamydial endometritis. peritonitis to doxycycline. Am J Obstet Gynecol 1977;129:614- J Clin Pathol 1985;38:726-32. 22. 6. Bell TA, James JF. Computer-assisted analysis of the therapy 19. Spence MR, Genadry R, Raffel L. Randomized prospective of acute salpingitis. Am J Obstet Gynecol 1980;138:1048-54. comparison of ampicillin and doxycycline in the treatment of 7. Thompson SE III, Hager WD, Wong KH, et al. The acute pelvic inflammatory disease in hospitalised patients. Sex microbiology and therapy of acute pelvic inflammatory disease Transm Dis 1981 ;8: 164-6. in hospitalized patients. Am J Obstet Gynecol 1980;136: 179-86. 20. Monif GRG. Clinical staging of acute bacterial salpingitis and 8. Hager WD, Eschenbach DA, Spence MR, Sweet RL. Criteria its therapeutic ramifications. Am J Obstet Gynecol for diagnosis and grading of salpingitis. Obstet Gynecol 1982; 143:489-95. 1983;61:113-4. 21. Eschenbach DA, Buchanan T, Pollock HM, et al. Poly- 9. McCormack WM, Nowroozi K, Alpert S, et al. Acute pelvic microbial etiology ofacute pelvic inflammatory disease. NEngi http://sti.bmj.com/ inflammatory disease. Characteristics ofpatients with gonococcal J Med 1975;293:166-71. on September 27, 2021 by guest. Protected copyright.