Acute Pelvic Inflammatory Disease and Clinical Response to

ANTUICROBiAL AGENT8 AND CtemoTmRmY, Feb. 1975, p. 133-138 Copyright 0 1975 American Society for Microbiology

Vol. 7, No. 2 Printed in U.S.A.

Acute Pelvic Inflammatory Disease and Clinical Response to Parenteral Doxycycline ANTHONY W. CHOW,* KAY L. MALKASIAN, JOHN R. MARSHALL, AND LUCIEN B. GUZE

Departments of Medicine, Obstetrics, and Gynecology, Harbor General Hospital, Torrance, California 90509,* Research and Medical Services, Veterans Administration Hospital (Wadsworth), Los Angeles, California 90073, and University of California School of Medicine, Los Angeles, California 90024 Received for publication 9 September 1974

The bacteriology of acute pelvic inflammatory disease (PID) and clinical to parenteral doxycycline were evaluated in 30 patients. Only 3 of 21 cul-de-sac cultures from PID patients were sterile, whereas all 8 normal control subjects yielded negative results (P< 0.005). Poor correlation was observed between cervical and cul-de-sac cultures. Neisseria gonorrhoeae, isolated from the cervix in 17 patients (57%), was recovered from the cul-de-sac only once. Streptococcus, Peptococcus, Peptostreptococcus, coliforms, and other organisms normally present in the vagina were the predominant isolates recovered from the cul-de-sac. Parenteral doxycycline resulted in rapid resolution of signs and symptoms (within 48 h) in 20 of 27 evaluable patients (74%). In five others, signs and symptoms of infection abated within 4 days. The remaining two patients failed to respond; in both cases, adnexal masses developed during doxycycline therapy. Gonococci were eradicated from the cervix in all but one patient who, nevertheless, had a rapid defervescence of symptoms. There was no clear-cut correlation between the clinical response and in vitro susceptibility of cul-de-sac isolates to doxycycline. These data confirm the usefulness of broad-spectrum antibiotics in acute PID. Culdocentesis is a reliable means of obtaining material for the bacteriological diagnosis of acute PID; however, the pathogenetic role and relative importance of gonococci and various other bacteria in acute PID need to be clarified further. response

Doxycycline, an analogue of tetracycline, is considerably more active in vitro than its parent compound against gonococci as well as a variety of other aerobic and anaerobic bacteria (11, 17, 18). Furthermore, it possesses pharmacological attributes, such as enhanced absorption and tissue penetration, that augment its in vivo effectiveness in experimental infections (5). Since both gonococci and diverse vaginal flora have been implicated either as primary or secondary pathogens in acute pelvic inflammatory disease (PIfD) (15), the efficacy of parenteral doxycycline in acute severe PIID was studied prospectively at Harbor General Hospital. Concurrently, the bacteriology of acute PIfD was investigated by comparative evaluation of cervical, cul-de-sac, and blood cultures obtained prior to antibiotic therapy. (This work was presented in part at the Vibramycin Symposium, San Francisco, Calif., 23 August 1974.)

and July 1973. The diagnosis of acute PIID was established by the presence of at least three of the following clinical criteria: fever greater than 100.4 F (ca. 38 C) (28 patients), purulent vaginal discharge (23 patients), pelvic tenderness (30 patients), peritoneal signs with rebound tenderness or ileus (27 patients), and pelvic mass (4 patients). Only one patient had received antibiotics prior to hospital admission (four doses of oral sulfisoxazole [Gantrisin ] for presumed urinary tract infection). Pregnant women and patients with known toxic or allergic reactions to tetracyclines were excluded from the

study. Cultures and susceptibility testing. Blood and cervical cultures were obtained prior to, during, and immediately after cessation of doxycycline therapy. Blood cultures were inoculated into aerobic brain heart infusion broth (Difco) as well as into prereduced, anaerobically sterilized brain heart infusion broth supplemented with yeast extract, cysteine, menadione, and hemin (Scott Laboratories). Blood cultures were periodically examined by Gram stain and routinely subcultured onto blood agar and prereduced, anaerobically sterilized media at 48 h. Cervical cultures were obtained under direct observation MATERIALS AND METHODS during speculum examination. In 21 patients, cul-dePatients. All 30 patients studied were hospitalized sac cultures were also obtained. Culdocentesis was at Harbor General Hospital between October 1972 performed with an 18-gauge spinal needle and syringe 133

134

ANTIMICROB. AGENTS CHEMOTHER.

CHOW ET AL.

after vaginal preparation with a 1:750 aqueous solution of benzalkonium chloride (Zephrin, Winthrop Laboratories). If the specimen was not obtained immediately, the cul-de-sac was irrigated with 2 to 5 ml of non-bacteriostatic, sterile physiological saline solution without preservative and reaspirated for culture. In addition, cul-de-sac cultures were similarly obtained in eight normal subjects free from pelvic disease, who were admitted to the hospital primarily for tubal ligation. Both cervical swabs and cul-de-sac specimens were transported in gassed containers to preclude exposure to air (Anaport or Anaswab, Scott Laboratories), processed immediately in the Infectious Disease Research Laboratory, and inoculated on routine and pre-reduced, anaerobically sterilized media by the Virginia Polytechnic Institute roll-tube technique (10). In addition, cervical cultures were inoculated on Thayer-Martin medium to preclude overgrowth of gonococci by vaginal flora; cul-desac aspirates were not routinely inoculated in ThayerMartin medium. All cultures were incubated aerobically, anaerobically, and under 10% CO2 (candle jar). Aerobic isolates were identified by standard procedures; obligate anaerobes were identified and speciated by the method of Holdeman and Moore (10). A Beckman pH meter was used for recording carbohydrate fermentation reactions, and a Dohrmann AnaBac chromatograph was used for fatty acid analysis. Antibiotic susceptibility testing of aerobic isolates was by the Kirby-Bauer disk diffusion method (2). Susceptibility of anaerobic isolates was determined by an agar dilution technique described previously (3). Therapeutic regimen and evaluation of response. A 200-mg dose of doxycycline was administered by intravenous infusion over 30 min on the day of admission, followed by 100 mg intravenously every 12 h thereafter. All patients were observed closely for clinical response and possible adverse reactions. Serial determinations of the hematocrit, leukocyte count, urinalysis, blood urea nitrogen, serum creatinine, serum bilirubin, alkaline phosphatase, glutamic oxalacetic transaminase, and lactic dehydrogenase were performed before and during therapy. Clinical response to therapy was considered "excellent" if systemic and local symptoms and signs resolved within 48 h, "good" if they resolved within 4 days, and "poor" if no significant clinical response was observed. RESULTS Bacteriology of acute PID. Blood cultures in

all 30 patients were uniformly negative. Cervical cultures yielded a total of 134 organisms, with an average of 4.5 isolates per specimen (Tables 1 and 2). Neisseria gonorrhoeae was isolated in 17 patients (57%). Cul-de-sac cultures were positive in 18 of 21 patients (Table 3), whereas all 8 normal control subjects yielded negative results (P < 0.005, x2 analysis). Fortyeight isolates were recovered from cul-de-sac cultures (Table 4). N. gonorrhoeae was isolated

TABLE 1. Aerotolerance of organisms in the cervical flora of 30 patients with acute PID No. of

No. with gonococci

18 11 1

11 6

patients Mixed aerobes and anaerobes ............ Aerobes only ........... No growth ..............I

TABLE 2. Relative incidence of various organisms in the cervical flora of 30 patients with acute PIDa No. isolated

% of total

Aerobic Streptococcus Group A Group D Staphylococcus sp. S. epidermidis S. aureus Coliforms Escherichia coli Proteus Neisseria gonorrhoeae Diphtheroids Lactobacillus Miscellaneous

104 29 4 8 23 20 3 16 10 4 17 10

78 22 3 6 17 15

6

4

3

2

Anaerobic Bacteroides sp. B. melaninogenicus B. capillosus B. fragilis Fusobacterium Peptococcus and Peptostreptococ-

30 9 4 2 1

13

22 7 3 1 1 1 10

5

4

2

1

Organism

1

2 12

7 3 13 7

cus

Veillonella and Acidaminococcus Gram-positive bacilli '

Cervical culture was sterile in one patient.

TABLE 3. Aerotolerance of organisms in cul-de-sac cultures of 21 patients with acute PID and 8 normal subjects Organism Organism

Aerobes only .......... Mixed aerobes and anaerobes .......... Anaerobes only ....... Cultures negative ..... a

Acute PID

~~(no.)

Controls

8

0

7 3 3a

0 0 8a

(no.)

Difference significant by x2 analysis (P < 0.005).

DOXYCYCLINE FOR ACUTE PD

VOL. 7, 1975

135

only once, even though simultaneous cervical cultures were positive for this organism in 13 of 21 patients. No qualitative or quantitative differences were noted when the cul-de-sac flora of gonococcal and non-gonococcal patients were compared. Bacteroidaceae were also notably absent in cul-de-sac specimens, whereas Streptococcus, Peptococcus; Peptostreptococcus, coliforms, and other organisms normally present in the vagina were the predominant isolates. There was poor correlation between cul-de-sac

and simultaneous cervical cultures. Only 31% of cul-de-sac isolates were cultured simultaneously from cervical specimens; similarly, only 19% of the cervical isolates were also recovered by culdocentesis. Antibiotic susceptibility. Since poor correlation between cervical and cul-de-sac cultures was observed and culdocentesis appeared to be a reliable technique for obtaining meaningful material for bacteriological diagnosis in acute PID, only the susceptibility data of cul-de-sac isolates are presented in Table 5. Eleven of 17 TABLE 4. Relative incidence of various organisms in aerobic isolates tested were susceptible to doxycul-de-sac cultures of 21 patients with acute PIDa cycline, whereas only 4 of 13 anaerobic isolates tested were inhibited by 2.5 gg of doxycycline % of total No. isolated Organism per ml. Aerobic 32 (12)b 67 Clinical response to parenteral doxycy13 (4) Streptococcus 27 cline. The clinical response to parenteral doxy5 (2) 10 Coliforms cycline therapy could be evaluated in 27 pa5 (1) 10 Diphtheroids tients (mean duration of therapy was 3 days). In 4 (4) 8 Staphylococcus the three remaining patients, severe side effects 4 Lactobacillus 2 developed which necessitated immediate disGonococcus 2 1(1) continuation of doxycycline therapy, and other 2 (0) 4 Candida antibiotics were administered. Despite general Anaerobic 16 (3) 33 lack of susceptibility in vitro by anaerobic Peptococcus and 12 (2) 25 isolates in particular, 20 patients (74%) had an Peptostreptococcus excellent response with rapid resolution of 3 (1) 6 Gram-positive bacilli symptoms and signs within 48 h. Five addiVeillonella 1 2 tional patients (19%) had a good response with resolution of symptoms and signs within 4 days. a Cultures were sterile in three patients. bNumber also recovered from simultaneous cervi- Only two patients failed to respond to doxycycal cultures. cline therapy. In both patients, adnexal masses TABLE 5. Susceptibility of cul-de-sac isolates to doxycycline Organisms

No. tested

No.tible til3e0k,g to3sk~

Streptococcus, group D Streptococcus, not group D Staphylococcus epidermidis

3 4 2

0 3 2

Escherichia coli Proteus mirabilis Diphtheroids Neisseria gonorrhoeae

1 2 4 1

0 4 1

No. with MICa (,Ug/Ml) Of: < 062

1.25

2.5

5

1

1

10

>_20

1 1

1 1 1

Aerobic

1

Anaerobic

Peptostreptococcus anaerobius P. micros

Peptococcus magnus P. asaccharolyticus Clostridium paraputrificum Lactobacillus minutis Eubacterium lentum Cumulative no. susceptible a

Minimal inhibitory concentration.

3 2 3 2

1

1 1 1

1

1

1

1 1 2

2

4

6

9

13

136

CHOW ET AL.

developed during doxycycline therapy, and peritoneal signs persisted. Initial cul-de-sac cultures were negative in one patient and positive for multiple organisms including Streptococcus, Staphylococcus epidermidis, aerobic Lactobacillus, and Candida sp. in the other. In neither case were gonococci isolated from any site. In the first patient, doxycycline therapy was, nevertheless, continued orally for 14 days with gradual defervescence of symptoms and resolution of adnexal masses. The second patient eventually responded to parenteral gentamicin and clindamycin therapy. Among these 27 patients who could be evaluated, cul-de-sac culture data were available only in 19 patients. In general, presence of aerobic bacteria in the cul-de-sac was more likely to be associated with a rapid (within 48 h) clinical response (10 of 12 rapid and 3 of 7 slow responders), whereas presence of resistant organisms in the cul-de-sac was correlated with a slower clinical response (5 of 7 slow and 6 of 12 rapid responders). However, these differences were not statistically signi-ficant. Anaerobic bacteria were isolated with the same frequency in the cul-de-sac cultures of both rapid (6 of 12) and slow (4 of 7) responders; similarly, the incidence of gonococci isolated from the cervix was comparable in both groups (12 of 20 rapid and 4 of 7 slow responders). Of interest was the observation for one patient who had rapid defervescence of symptoms (within 48 h) after doxycycline therapy despite persistence of gonococci (susceptible to doxycycline in vitro) in the cervix. Initial cul-de-sac cultures of this patient grew a mixture of Peptococcus asaccharolyticus, Streptococcus, Acinetobacter, and S. epidermidis. Tolerance and toxicity. Parenteral doxycycline was generally well tolerated in the majority of our patients. Side effects were noted in five patients, in three of whom immediate discontinuation of medication was necessary. One patient experienced an acute reaction with nausea, chills, light-headedness, and jerking motions of upper extremities immediately after the antibiotic infusion. One patient developed a generalized urticarial rash. Three patients developed superficial thrombophlebitis. No hepatic or renal toxicity was encountered during doxycycline therapy. Effect of doxycycline on the cervical flora. In 23 patients, the effect of doxycycline on the cervical flora could be evaluated from pre- and post-therapy cultures (Table 6). Twenty-four of 116 initial cervical isolates (21%) persisted in the post-therapy specimens. N. gonorrhoeae was eradicated in all except one instance. Note-

ANTIMICROB. AGENTS CHEMOTHER. TABLE 6. Effect of doxycycline on the cervical flora in 23 patients with acute PID Post-therapy Pre-

Organism

therapy culture

culture

Persis-

Acqui-

tance'

sitionb

Aerobic Streptococcus Staphylococcus Coliforms Neisseria gonorrhoeae Diphtheroids Lactobacillus Candida species Miscellaneous

91 24 22 12 16 9 6 1

21 7 5 4 1

1

0

41 14 9 4 0 3 7 2 2

Anaerobic Bacteroides Fusobacterium Peptococcus and Peptostreptococcus Ruminococcus Veillonella and Acidaminococcus Nonsporulating grampositive bacilli

25 7 1 11

3 3 0 0

48 16 0 14

0 4

0 0

2 4

2

0

12

2

2 0

a Presence of the same organism in both pre- and post-therapy cultures. 'Presence of organism only in post-therapy cultures.

worthy was the finding that, whereas aerobic isolates diminished by 32%, anaerobic isolates increased by 104% in the post-therapy cultures. Increased colonization of the cervical flora was particularly striking for Bacteroides, anaerobic nonsporeforming gram-positive bacilli, and aerobic Lactobacillus. DISCUSSION Several aspects of this study deserve comment. First, culdocentesis appears to be a safe and reliable way of obtaining meaningful material for specific microbiological diagnosis in acute PID. The consistently negative cul-de-sac cultures from all eight normal control subjects suggest that contamination by vaginal flora should not be a problem provided that culdocentesis is performed properly. The extremely poor correlation observed between cervical and cul-de-sac cultures further substantiates the value of this procedure; others have similarly noted the diagnostic usefulness of cul-de-sac cultures in recurrent PID (19). It should be noted, however, that all except one of our patients in whom culdocentesis was performed had associated peritoneal signs with rebound

VOL. 7, 1975

tenderness or ileus, which may account for the high recovery rate. The primary and probably only major risk with this procedure is that of rupturing an unsuspected tubo-ovarian abscess into the peritoneal cavity. This can be avoided by a careful bimanual pelvic examination to exclude patients with large masses not affixed to the cul-de-sac. Second, the paucity of gonococci in cul-de-sac cultures, despite recovery from the cervix, is intriguing. Since a selective medium was not routinely used in cul-de-sac cultures, it is possible that recovery of some gonococci may have been hampered due to our study design. It is questionable whether exposure to saline during culdocentesis and specimen transport could have interfered with the isolation of gonococci, since sterile physiological saline rather than bacteriostatic saline was used. Similarly, our method of specimen transport should not cause a deleterious effect since gonococci survive well in the Anaport system for at least 24 h (unpublished observation). Nevertheless, this paucity of gonococcal isolation from the cul-de-sac or peritoneal exudates, despite recovery from the cervix of PID patients, has also been noted by a number of investigators (7, 12). These data and the nature of aerobic and anaerobic bacteria isolated from the cul-de-sac of our patients would suggest that, whereas the gonococcus may be important in initiating acute PID, its primary role appears to be paving the way for secondary invaders from the normal vaginal flora to gain access to the upper genital tract. Third, the excellent clinical response of acute PID to parenteral doxycycline therapy observed in our patients is in keeping with clinical observations by others that broad-spectrum antibiotics such as ampicillin or tetracyclines are particularly useful in the treatment of acute PID (4, 15). Doxycycline differs appreciably from its parent tetracycline in that it is rapidly and almost completely absorbed by the oral route and has a prolonged serum half-life (13). After administration of 200-mg therapeutic doses daily, serum concentrations reportedly range from 2 to 3.5 ug/ml (13). It can be administered in the presence of renal insufficiency without moderation of normal dosage (14). It has optimal lipophilicity and penetrates efficiently into tissues (16). It is considerably more active in vitro than tetracycline against a variety of both aerobic and anaerobic bacteria (11, 17, 18). The lack of better correlation between in vitro susceptibility of individual cul-de-sac isolates and the clinical response to doxycycline is

DOXYCYCLINE FOR ACUTE PID

137

intriguing. One possible explanation is that tissue concentrations of doxycycline in pelvic organs may be considerably higher than achievable serum concentrations, as has been documented in the kidney, lung, and prostate (6, 8). Furthermore, in mixed infections due to multiple organisms, as often is the case in acute PID, antibiotic susceptibility of individual components of a complex flora may not be as important as the net effect of the antibiotic on the total microbial ecology. That this may be true is suggested by recent observations that intraabdominal and pleuropulmonary infections due to complex microbial flora may be effectively treated by antibiotics only active against certain components of the flora and despite presence of resistant organisms (1, 9). This is an extremely complex problem which undoubtedly deserves further investigation. ACKNOWLEDGMENTS We are grateful to Patrick Cunningham, Valerie Patten, Jeanine Mettee, and Ane Pennington for technical assistance, and to Lorraine Fugita for preparation of this manuscript. This work was supported by a grant from Pfizer Laboratories. LITERATURE CITED 1. Bartlett, J. G., and S. M. Finegold. 1974. Anaerobic infections of the lung and pleural space. Am. Rev. Respir. Dis. 110:56-77. 2. Bauer, A. W., W. M. M. Kirby, J. C. Sherris, and M. Turck. 1966. Antibiotic susceptibility testing by a standardized single disc method. Am. J. Clin. Pathol. 45:493-496. 3. Chow, A. W., J. Z. Montgomerie, and L. B. Guze. 1974. Parenteral clindamycin for severe anaerobic infections. Arch. Intern. Med. 143:78-82. 4. Durfee, R. B. 1972. Infections of the female genital tract, p. 475-498. In P. D. Hoeprich (ed.), Infectious diseases. Harper & Row Publishers, Hagerstown. 5. English, A. R. 1966. Alpha-6-deoxyoxytetracycline. I. Some biological properties. Proc. Soc. Exp. Biol. Med. 122:1107-1112. 6. Fabre, J., E. Milek, P. Kalfopoulos, and G. Merier. 1971. Behavior of tetracyclines in man. II. Excretion, penetration into normal and inflammatory tissue, behavior in renal insufficiency and hemodialysis. Schweiz. Med. Wochenschr. 101:625-633. 7. Falk, V. 1965. Treatment of acute non-tuberculous salpingitis with antibiotics alone and in combination with glucocorticoids. A prospective double blind controlled study of the clinical course and prognosis. Acta Obstet. Gynecol. Scand. 44(Suppl. 6):1-118. 8. Garnes, H. A. 1973. Doxycycline levels in serum and prostatic tissue in man. Urology 1:205-207. 9. Gorbach, S. L., and H. Thadepalli. 1974. Clindamycin in pure and mixed anaerobic infections. Arch. Intern. Med. 134:87-92. 10. Holdeman, L. V., and W. E. C. Moore (ed.). 1972. Anaerobe laboratory manual. Yirginia Polytechnic Institute Anaerobe Laboratory, Blacksburg. 11. Isenberg, H. D. 1967. In vitro activity of doxycycline against bacteria from clinical material. Appl. Microbiol. 15:1074-1078. 12. Jacobson, L. 1964. Laparoscopy in the diagnosis of acute

138

CHOW ET AL.

salpingitis. Acta Obstet. Gynecol. Scand. 43:160-175. 13. Leibowitz, B. J., J. L. Hakes, M. M. Cahn, and E. J. Levy. 1972. Doxycycline blood levels in normal subjects after intravenous and oral administration. Curr. Ther. Res. Clin. Exp. 14:820-831. 14. Mahon, W. A., J. V. P. Wittenberg, and P. G. Tuffnel. 1970. Studies on the absorption and distribution of doxycycline in normal patients and in patients with severely impaired renal function. J. Can. Med. Assoc. 103:1031-1034. 15. Mead, P. B., and D. B. Louria. 1969. Antibiotics in pelvic infections. Clin. Obstet. Gynecol. 12:219-239. 16. Schach von Wittenau, M., and R. Yeary. 1963. The

ANTIMICROB. AGENTS CHEMOTHER. excretion and distribution in body fluids of tetracyclines after intravenous administration to dogs. J. Pharmacol. Exp. Ther. 140:258-266. 17. Steigbigel, N. G., C. W. Reed, and M. Finland. 1968. Susceptibility of common pathogenic bacteria to seven tetracycline antibiotics in vitro. Am. J. Med. Sci. 255:179-195. 18. Sutter, V. L., F. P. Tally, Y.-Y. Kwok, and S. M. Finegold. 1973. Activity of doxycycline and tetracycline versus anaerobic bacteria. Clin. Med. 80:31-38. 19. Swenson, R. M., T. C. Michaelson, M. J. Daly, and E. H. Spaulding. 1973. Anaerobic bacterial infections of the female genital tract. Obstet. Gynecol. 42:538-541.