Home Sampling versus Conventional Swab ... - Semantic Scholar

951

Home Sampling versus Conventional Swab Sampling for Screening of Chlamydia trachomatis in Women: A Cluster-Randomized 1-Year Follow-up Study Lars Østergaard,1 Berit Andersen,3 Jens K. Møller,2 and Frede Olesen3

Departments of 1Infectious Diseases and 2Clinical Microbiology, Aarhus University Hospital, and 3Research Unit and Department of General Practice, University of Aarhus, Aarhus, Denmark

We compared the efficacy of a screening program for urogenital Chlamydia trachomatis infections based on home sampling with that of a screening program based on conventional swab sampling performed at a physician’s office. Female subjects, comprising students at 17 high schools in the county of Aarhus, Denmark, were divided into a study group (tested by home sampling) and a control group (tested in a physician’s office). We assessed the number of new infections and the number of subjects who reported being treated for pelvic inflammatory disease (PID) at 1 year of follow-up; 443 (51.1%) of 867 women in the intervention group and 487 (58.5%) of 833 women in the control group were available for follow-up. Thirteen (2.9%) and 32 (6.6%) new infections were identified in the intervention group and the control group, respectively (Wilcoxon exact value, P p .026 ). Nine (2.1%) women in the intervention group and 20 (4.2%) in the control group reported being treated for PID (P p .045), indicating that a screening strategy involving home sampling is associated with a lower prevalence of C. trachomatis and a lower proportion of reported cases of PID.

Urogenital Chlamydia trachomatis infection is a common sexually transmitted infection that primarily affects adolescents [1, 2]. In women, the infection is asymptomatic in more than half the cases [3]. Despite this, C. trachomatis may cause pelvic inflammatory disease (PID) [4], which may result in chronic abdominal pain [5], female infertility [6], or ectopic pregnancy [7–9]. If recognized, however, the infection can be treated easily and effectively with azithromycin [10]. Until recently, the diagnosis of urogenital C. trachomatis infections in women could only be obtained by testing urethral and endocervical swab samples taken by a health care professional. With the introduction of DNA amplification techniques, however, self-collectible samples, such as urine, vaginal swabs, and vaginal flushes, have been used with the same diagnostic efficacy as swab samples obtained by health care professionals [11–18]. This also includes samples taken at home and directly

Received 18 October 1999; revised 15 March 2000; electronically published 25 October 2000. Informed consent was obtained from all subjects. The Helsinki Declaration and the Guidelines for Good Clinical Practice were followed. The study was approved by the Central Danish Ethical Committee and the Danish Data Registration Agency. Financial support: Danish National Board of Health (grant 210, 1997), Løvens Kemiske Fabriks Research Foundation, Nycomed DAK, and Chairman Jacob Madsen’s & Hustru Olga Madsen’s Foundation, Helga and Peter Kornings Foundation, and Lægekredsforeningen i Aarhus Amt (the Aarhus County Medical District Association). Reprints or correspondence: Dr. Lars Østergaard, Dept. of Infectious Diseases, Aarhus University Hospital, P.P. Ørumsgade 11, DK-8000 Aarhus C, Denmark ([email protected]). Clinical Infectious Diseases 2000; 31:951–7 q 2000 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2000/3104-0016$03.00

mailed to the laboratory [19, 20]. Clearly, such strategies might be used for screening programs in outreach settings [21, 22]. In a cluster-randomized setting, we previously showed that a screening strategy involving home sampling and mailed samples increased the number of tested women 11-fold when compared with a strategy that used swab samples obtained by doctors [23]. Furthermore, the use of home sampling has improved the identification of infected partners when compared with conventional methods of partner tracing [24]. Thus home sampling is an efficient method for intensified testing of apparently healthy individuals who do not seek a health care professional. It therefore also may reduce the prevalence of disease and its complications over time. To our knowledge, these outcome measures have not yet been assessed in clinical studies. The objective of this study was to compare a screening strategy based on home sampling with a strategy of conventional testing in order to determine the prevalence of disease after 1 year and the number of treated PID cases during the 1 year of follow-up.

Patients and Methods Study population and randomization. A flow chart of the study is shown in figure 1. All 17 high schools in the county of Aarhus, Denmark, were cluster randomized 1:1 by simple redeeming (drawing lots from a hat) into an intervention (home sampling) group consisting of 8 high schools composed 2603 women and 1733 men, and a control group consisting of 9 high schools composed 2884 women and 1689 men. More than 95% of the population was white and Protestant. Intervention. From January 1997 to April 1997, the 2603

952

Østergaard et al.

Figure 1.

CID 2000;31 (October)

CONSORT flow chart for the 2 arms of the study. PID, pelvic inflammatory disease.

women and 1733 men in the home sampling group all received a test kit for home sampling [19, 23]. At baseline, the home sampling kits were given to the students at the end of gatherings at which information about the C. trachomatis diseases and the study was given. The home sampling kit consisted of a vaginal pipette (containing 5 mL sterile sodium chloride) for obtaining vaginal flush samples (women) and a urine sample (men), a questionnaire, written instructions on how to obtain the sample, and a self-addressed, stamped envelope. Students were instructed to administer the vaginal pipette for sampling on receipt. The samples obtained at home were mailed by the students directly to the Department of Clinical Microbiology, where they were analyzed. The students also provided the address where the test results were to be sent. Students with positive test results were requested in writing to visit a doctor for treatment and partner tracing and to take a letter to the doctor. The letter contained information about the study and a slip for the doctor to fax back to the Department of Clinical Microbiology in order to confirm that treatment had been given. The intervention group only had the ability to be tested by home sampling once until follow-up after 1 year, when a new home-obtained sample

was requested in order to assess the prevalence of C. trachomatis infection at that time. Control. The control group received the same information and questionnaire as the home sampling group, but they were not supplied with the home sampling kit. Instead, they were offered a free testing at the local clinic for sexually transmitted diseases (STDs) or at the office of any other physician, including that of their general practitioner. Students were informed of the possibility of being tested right after receipt of the written information on the C. trachomatis infection. All samples from STD clinics and general practitioners in the county of Aarhus are analyzed at the Department of Clinical Microbiology, and it was therefore possible by use of the students’ unique central registration numbers to assess the number of tested and infected students in the control group at baseline. Follow-up. Although we provided information to both men and women, we only intended to follow up on the women, because serious complications primarily affect this sex. One year after the initial testing (January 1998–April 1998), the participants who responded to the questionnaire at baseline and who also agreed to


Home Sampling for Diagnosis of C. trachomatis

be contacted 1 year later (867 women in the home sampling group and 833 in the control group) received a home sampling kit that was directly mailed to their home address. This current address was obtained from the Central Personnel Register in Denmark, in which addresses of all Danish citizens are continually updated. The home sampling kit for follow-up consisted of a vaginal pipette, written instructions, and a questionnaire. The women received 1 reminder by mail 1 month after the first follow-up material was sent out. The vaginal sample collected at home and the completed questionnaire were mailed directly to the Department of Clinical Microbiology. Test results, treatment, and partner tracing were dealt with as described for baseline. The questionnaire included information about treatment for PID and admittance to a hospital for PID during the year of follow-up. In an attempt to verify that treatment for PID had been given, every student who reported treatment for PID was sought among all records of antimicrobial prescriptions at the central Danish register for prescriptions (Lægemiddelstyrelsen). Outcome measures. The outcome measures were the following: prevalence of C. trachomatis infections after 1 year of follow-up; the proportion of women reporting treatment for PID; and the proportion of women admitted to hospital for PID. The minimum important differences for each of the outcome measures were P ! .05. Control for selection bias and confounding. The questionnaire at baseline also required the subjects to provide information about age, any regular intimate relationship, urogenital symptoms, and 4 questions addressing the knowledge of C. trachomatis infection. All microbiological testing for C. trachomatis and Neisseria gonorrhoeae in the county of Aarhus are centralized to the Department of Clinical Microbiology, Aarhus University Hospital, and all tests, the test results, and the setting for sampling are stored in a database in the department. It was therefore possible by use of the participants’ unique central registration numbers to assess the number of respondents at follow-up who in the period 1993–1997 (up to baseline) were tested for C. trachomatis, N. gonorrhoeae, or both, the results of the test, and the setting (general practitioner, STD clinic, or hospital). These data were obtained to illuminate the health care utilization in the 2 groups with regard to STDs. The parameters above were compared between the intervention and control groups after 1 year of follow-up. By doing so, we could assess the implication of selection bias due to loss to follow-up (table 1). Statistical methods. This is a cluster-randomized study, and the unit of randomization and, therefore, the unit of analysis, were the high school. Thus, in order to test for significance, incidence proportions (number of events divided by number of respondents) were calculated for every high school with regard to number of new C. trachomatis infections and number of women with PID (in total and admitted to a hospital) [25]. For every outcome measure, a mean incidence proportion for each of the 2 groups of high schools (intervention group and control group) was calculated. For categorical data that were not normally distributed, the Wilcoxon exact P value was calculated to test for significance between the mean incidence proportions of the 2 groups of high schools. Furthermore, the difference between the mean incidence proportions and the corresponding 95% confidence interval (CI) was calculated by means of Student’s t test.

953

The comparison of demographic, clinical, and behavioral data between the respondents at follow-up were assessed by x2 test or Fisher’s exact test, as appropriate, in order to assess a difference in selection bias and confounding between the 2 groups. The statistical calculations were performed by SPSS 9.0 statistical software (Chicago, IL). Microbiological analyses. The vaginal pipette contained 5 mL sterile, isotonic sodium chloride. The woman introduced the tip of the pipette into the vagina, which was flushed by the contents of the pipette. Next, the saline was aspirated into the pipette, which was then mailed by ordinary mail to the laboratory, where it was analyzed within 3 days of sampling. The vaginal flush samples were treated according to the manufacturer’s protocol for urine samples and were tested by a transcription-mediated amplification assay (AMP Ct, Gen-Probe; San Diego, CA). Patients were designated infected only if the result obtained by the AMP Ct could be confirmed by the ligase chain reaction (LCx, Abbott Diagnostics; Abbott Park, IL).

Results Test rate and prevalence at baseline in the 2 groups. At baseline, a total of 1254 (48.2%) women in the home sampling group and 1097 (38.0%) in the control group responded by completing and returning the questionnaire. Of these, 928 in the home sampling group and 833 in the control group were sexually experienced. Of these, 867 and 63 were tested in the home sampling group and control group, respectively, and 43 infections were detected in the home sampling group, compared with 5 infections in the control group. The difference in test rate and number of detected chlamydia cases between the intervention group and the control group are highly statistically significant, as described elsewhere [23]. A total of 867 women in the home sampling group and 833 women in the control group agreed to take part in the followup study. Similarity between the follow-up respondents. Of the 867 women in the home sampling group who agreed to be followed up, 443 responded to the follow-up request after 1 year (51.1%), and, of the 833 women in the control group, 487 responded (58.5%). In table 1, demographic data, clinical data, and health care utilization data are given for respondents in the 2 groups. There was no difference between the 2 groups with regard to age distribution, knowledge of C. trachomatis infection, current regular intimate relationship, urogenital symptoms, and health care utilization with regard to STDs during the previous 5 years. Outcome after 1 year of follow-up. Thirteen new infections were found in the intervention group (2.9%), compared with 32 (6.6%) in the control group (Wilcoxon exact value, P p .026). A total of 9 students (2.1%) were treated for PID in the intervention group, compared with 20 (4.2%) in the control group (Wilcoxon exact value, P p .045). Although not statistically significant, there was some indication that fewer women in the home sampling group were admitted to a hospital for

954

Østergaard et al.

Table 1.


Similarity between the 2 groups that responded to follow-up.

Parameter Age, years 15 16 17 18 >19 Knowledge of Chlamydia trachomatis infection Chlamydia is an STD Yes No Not answered Chlamydia may be asymptomatic Yes No Not answered Protection through condom use Yes No Not answered Chlamydia may be cured Yes No Not answered Regular intimate relationship Yes No Not answered Any urogenital symptoms Yes No Not answered Health care utilization in the previous 5 years C. trachomatis test at GP Yes No C. trachomatis test at STD clinic Yes No Infected with C. trachomatis Yes No Gonorrhea test at GP Yes No Gonorrhea test at STD clinic Yes No Infected with gonorrhea Yes No

Control group (n p 487)

Intervention group (n p 443)

P .12

37 131 158 111 50

(7.6) (26.9) (32.4) (22.8) (10.3)

44 116 148 97 38

(9.9) (26.2) (33.4) (21.9) (8.6) .40

485 (99.6) 1 (0.2) 1 (0.2)

442 (99.8) 0 (0.0) 1 (0.2)

482 (99.0) 4 (0.8) 1 (0.2)

436 (98.4) 3 (0.7) 4 (0.9)

486 (99.8) 1 (0.2) 0 (0.0)

442 (99.8) 1 (0.2) 0 (0.0)

452 (92.8) 33 (6.8) 2 (0.4)

410 (92.6) 28 (6.3) 5 (1.1)

293 (60.2) 193 (39.6) 1 (0.2)

281 (63.4) 161 (36.3) 1 (0.2)

70 (14.4) 416 (85.4) 1 (0.2)

76 (17.2) 362 (81.7) 5 (1.1)

157 (32.2) 330 (67.8)

144 (32.5) 299 (67.5)

2 (0.4) 485 (99.6)

1 (0.2) 442 (99.8)

16 (3.3) 471 (96.7)

12 (2.7) 431 (97.3)

49 (10.1) 438 (89.9)

39 (8.8) 404 (92.1)

2 (0.4) 485 (99.6)

1 (0.2) 442 (99.8)

0 (0.0) 487 (100)

0 (0.0) 443 (100)

.34

.95

.44

.59

.10

.93

1.00

.61

.51

1.00

—

NOTE. Data are no. (%) of respondents. x2 or Fisher’s exact test was used, as appropriate. GP, general practitioner; STD, sexually transmitted disease.

PID, compared with the conventional swab sampling group (Wilcoxon exact value, P p .127) (table 2). Testing during the follow-up period. During the 1-year follow-up period (after baseline testing and before the time of follow-up), 2 patients (0.5%) in the intervention group and 5 patients (1.0%) in the control group were tested for C. trachomatis at the STD clinic (not significant). In the same period, 127 (28.7%) patients in the intervention group and 171 (35.1%) in the control group had a C. trachomatis test performed by a general practitioner (x 2 p 4.13, P p .04). Thus, despite the

high number of tests performed in both groups, which might reflect the high degree of opportunistic screening among women in Denmark, it seems that offering the opportunity of home sampling may reduce the need for later testing at a physician’s office. Discussion In a 1-year follow-up study, we compared the efficacy of a school-based home sampling strategy for screening of C. tra-


Table 2.


955

Incidence proportions of infections and complications for students in the 2 groups of high schools in Aarhus, Denmark.

Group Control group 1 2 3 4 5 6 7 8 9 Not identifiable Total MIP, % Intervention group 10 11 12 13 14 15 16 17 Not identifiable Total MIP, % Difference of MIP, % Lower 95% CI for DM, % Higher 95% CI for DM, % Wilcoxon exact P

No. of students followed up

No. of infected students at follow-up

Incidence proportion for infections at follow-up, %

No. treated for PID

Incidence proportions for PID treatment, %

No. hospitalized for PID

Incidence proportions for hospitalized PID, %

65 22 27 50 69 48 79 46 70 11 487 —

1 4 4 4 1 3 2 2 10 1 32 —

1.54 18.18 14.81 8.00 1.45 6.25 2.53 4.35 14.29 9.09 — 8.05

2 4 0 2 2 3 2 2 2 1 20 —

3.08 18.18 0.00 4.00 2.90 6.25 2.53 4.35 2.86 9.09 — 5.32

0 1 0 1 1 1 0 0 1 0 5 —

0.00 4.55 0.00 2.00 1.45 2.08 0.00 0.00 1.43 0.00 — 1.15

54 46 48 49 51 69 43 73 10 443 — —

0 3 2 0 2 4 0 2 0 13 — —

0.00 6.52 4.17 0.00 3.92 5.80 0.00 2.74 0.00 — 2.57 5.48

1 2 2 1 1 2 0 0 0 9 — —

1.85 4.35 4.17 2.04 1.96 2.90 0.00 0.00 0.00 — 1.92 3.40

0 1 0 0 0 0 0 0 0 1 — —

0.00 2.17 0.00 0.00 0.00 0.00 0.00 0.00 0.00 — 0.24 0.91

—

—

0.95

—

20.38

—

20.23

— —

— —

10.00 .026

— —

7.19 .045

— —

2.05 .127

NOTE. The intervention group took samples at home and mailed them to be assessed. Members of the control group obtained conventional swab sampling at a physician’s office. The mean incident proportions for each outcome measure were compared between the intervention group and the control group. The Wilcoxon exact P value (2-tailed) was used for assessing statistical significance (P ! .05 ); 95% CI, 95% confidence interval; DM, difference of mean; MIP, mean incidence proportion; PID, pelvic inflammatory disease.

chomatis infection with that of a school-based screening offering conventional invasive swab sampling performed in a physician’s office. There was an association between the former screening strategy and a reduced prevalence of C. trachomatis infection after 1 year of follow-up and a lower number of PID cases in the period during follow-up. Studies on selective screening by swab sampling, in which testing is based on >1 risk factors, have indicated that these strategies may reduce the prevalence of C. trachomatis infection [26–28]. Scholes et al. [29] showed that the number of PID cases could be reduced by selective screening. However, up to twothirds of infected women are asymptomatic, and, therefore, a screening strategy that allows intensified testing of apparently healthy individuals who do not seek a health care provider is needed. A screening strategy based on home sampling as described here may increase access to health care, because sampling can be performed by the individual herself, in privacy, without unpleasant invasive sampling and at a convenient time, without loss of wages and without missing school. Several studies have shown that self-collected samples can be used for the detection of urogenital C. trachomatis infection both inside and outside clinical settings [22, 30, 31], but the

effect of this with regard to prevalence or clinical outcome measures has not been assessed. At baseline, a high number of women declined to respond to the study. Many of these might have been sexually inexperienced, and the study may therefore have seemed irrelevant to them. Also, almost half the study population was lost to followup. However, we know from comparison of demographics, clinical data, and health care utilization with regard to STDs during the previous 5 years that the 2 groups of respondents after 1 year of follow-up seem similar. Although we were not able to assess all aspects of behavior, it seems that selection bias had no major impact on the findings in the study. A C. trachomatis infection may last for years if undiscovered and untreated, and this was probably the reason why a single episode of home testing in this study’s intervention group could explain the reduced prevalence of infection at 1 year of followup: more infections in the home sampling group were simply discovered and treated at baseline compared with the control group. It is argued that PID that occurs after a C. trachomatis infection [3, 5, 32] precedes long-term complications such as ectopic pregnancy [7] and infertility [8]. These potential long-term

956

Østergaard et al.

complications, however, will first be noticed when the students in this study begin considering pregnancy. This could be many years in the future, and the presence of PID rather than infertility and ectopic pregnancy was therefore chosen as a measure of complication. We found that the risk of having had PID during the 1-year follow-up period was reduced in the home sampling group. The data regarding PID were obtained by self-reporting, which may have occasioned information bias. However, it is likely that the extent of this bias is equal in both groups. This is supported by the fact that almost identical rates (44% in the home sampling group and 43% in the control group) of the students who reported being treated for PID were centrally registered as having had a prescription on macrolides, tetracycline, pivampicillin, or metronidazole. Patients who were treated in general practice with sample packages of antibiotics that were handed out are not centrally registered, and therefore they could not be traced. The number of PID cases may have been underestimated in both groups because patients with silent PID would not have been treated or admitted to a hospital. The diagnosis, however, is highly difficult to obtain without laparoscopy [33, 34], which would not have been acceptable to all the enrolled students. We conclude that a screening strategy involving home sampling is associated with a lower prevalence of C. trachomatis at 1 year of follow-up and with a lower proportion of reported cases of PID when compared with a conventional screening strategy based on swab sampling in a physician’s office. A broad-based home sampling screening strategy might contribute to lowering the prevalence of chlamydia and to lowering the risk of PID for the individual. Acknowledgments We gratefully acknowledge the technical assistance of Jonna Guldberg, Gitte Høj, and Mette Jensen.

References 1. Shafer MA, Schachter J, Moncada J, et al. Evaluation of urine-based screening strategies to detect Chlamydia trachomatis among sexually active asymptomatic young males. JAMA 1993; 270:2065–70. 2. Burstein GR, Waterfield G, Joffe A, Zenilman JM, Quinn TC, Gaydos CA. Screening for gonorrhoeae and chlamydia by DNA amplification in adolescents attending middle school health centers: opportunity for early intervention. Sex Transm Dis 1998; 25:395–402. 3. Cates WJ, Wasserheit JN. Genital chlamydial infections: epidemiology and reproductive sequelae. Am J Obstet Gynecol 1991; 164:1771–81. 4. Westro¨m L, Wo¨lner-Hanssen P. Pathogenesis of pelvic inflammatory disease. Genitourin Med 1993; 69:9–17. 5. Stacey C, Munday P, Thomas B, Gilchrist C, Taylor-Robinson D, Beard R. Chlamydia trachomatis in the fallopian tubes of women without laparoscopic evidence of salpingitis. Lancet 1990; 336:960–3. 6. Tubal infertility: serologic relationship to past chlamydial and gonococcal infection. World Health Organization Task Force on the Prevention and Management of Infertility. Sex Transm Dis 1995; 22:71–7.


7. Brunham RC, Peeling R, Maclean I, Kosseim ML, Paraskevas M. Chlamydia trachomatis–associated ectopic pregnancy: serologic and histologic correlates. J Infect Dis 1992; 165:1076–81. 8. Paavonen J, Vesterinen E, Aantaa K, Rasanen J. Factors predicting abnormal hysterosalpingographic findings in patients treated for acute pelvic inflammatory disease. Int J Gynaecol Obstet 1985; 23:171–5. 9. Egger M, Low N, Smith GD, Lindblom B, Herrmann B. Screening for chlamydial infections and the risk of ectopic pregnancy in a county in Sweden: ecological analysis. BMJ 1998; 316:1776–80. 10. Martin DH, Mroczkowski TF, Dalu ZA, et al. A controlled trial of a single dose of azithromycin for the treatment of chlamydial urethritis and cervicitis. The Azithromycin for Chlamydial Infections Study Group. N Engl J Med 1992; 327:921–5. 11. Grun L, Tassano-Smith J, Carder C, et al. Comparison of two methods of screening for genital chlamydial infection in women attending in general practice: cross sectional survey. BMJ 1997; 315:226–30. 12. Wiesenfeld HC, Heine RP, Rideout A, Macio I, DiBiasi F, Sweet RL. The vaginal introitus: a novel site for Chlamydia trachomatis testing in women. Am J Obstet Gynecol 1996; 174:1542–6. 13. Hook EW, Smith K, Mullen C, et al. Diagnosis of genitourinary Chlamydia trachomatis infections by using the ligase chain reaction on patientobtained vaginal swabs. J Clin Microbiol 1997; 35:2133–5. 14. Schachter J, Stamm WE, Quinn TC, Andrews WW, Burczak JD, Lee HH. Ligase chain reaction to detect Chlamydia trachomatis infection of the cervix. J Clin Microbiol 1994; 32:2540–3. 15. Chernesky MA, Lee H, Schachter J, et al. Diagnosis of Chlamydia trachomatis urethral infection in symptomatic and asymptomatic men by testing firstvoid urine in a ligase chain reaction assay. J Infect Dis 1994; 170:1308–11. 16. Stary A, Tomazic-Allen S, Choueiri B, Burczak J, Steyrer K, Lee H. Comparison of DNA amplification methods for the detection of Chlamydia trachomatis in first-void urine from asymptomatic military recruits. Sex Transm Dis 1996; 23:97–102. 17. Lee HH, Chernesky MA, Schachter J, et al. Diagnosis of Chlamydia trachomatis genitourinary infection in women by ligase chain reaction assay of urine. Lancet 1995; 345:213–6. 18. Schachter J, Moncada J, Whidden R, et al. Noninvasive tests for diagnosis of Chlamydia trachomatis infection: application of ligase chain reaction to first-catch urine specimens of women. J Infect Dis 1995; 172:1411–4. 19. Østergaard L, Møller JK, Andersen B, Olesen F. Diagnosis of urogenital Chlamydia trachomatis infection in women based on mailed samples obtained at home: multipractice comparative study. BMJ 1996; 313:1186–9. 20. Morre SA, Van VI, de JA, et al. Mailed, home-obtained urine specimens: a reliable screening approach for detecting asymptomatic Chlamydia trachomatis infections. J Clin Microbiol 1999; 37:976–80. 21. Cook RL, St George K, Lassak M, Tran N, Anhalt JP, Rinaldo CR. Screening for Chlamydia trachomatis infection in college women with a polymerase chain reaction assay. Clin Infect Dis 1999; 28:1002–7. 22. Gaydos CA, Howell MR, Pare B, et al. Chlamydia trachomatis infections in female military recruits. N Engl J Med 1998; 339:739–44. 23. Østergaard L, Andersen B, Olesen F, Møller JK. Efficacy of home sampling for screening of Chlamydia trachomatis: randomised study. BMJ 1998; 317:26–7. 24. Andersen B, Østergaard L, Møller JK, Olesen F. Home sampling versus conventional contact tracing for detecting Chlamydia trachomatis infection in male partners of infected women: randomised study. BMJ 1998; 316: 350–1. 25. Kerry SM, Bland JM. Analysis of a trial randomised in clusters. BMJ 1998; 316:54. 26. Mosure DJ, Berman S, Fine D, DeLisle S, Cates WJ, Boring JR. Genital Chlamydia infections in sexually active female adolescents: do we really need to screen everyone? J Adolesc Health 1997; 20:6–13. 27. Addiss DG, Vaughn ML, Ludka D, Pfister J, Davis JP. Decreased prevalence of Chlamydia trachomatis infection associated with a selective screening



program in family planning clinics in Wisconsin. Sex Transm Dis 1993; 20:28–35. 28. Katz BP, Blythe MJ, Van der Pol B, Jones RB. Declining prevalence of chlamydial infection among adolescent girls. Sex Transm Dis 1996; 23: 226–9. 29. Scholes D, Stergachis A, Heidrich FE, Andrilla H, Holmes KK, Stamm WE. Prevention of pelvic inflammatory disease by screening for cervical chlamydial infection. N Engl J Med 1996; 334:1362–6. 30. Marrazzo JM, White CL, Krekeler B, et al. Community-based urine screening for Chlamydia trachomatis with a ligase chain reaction assay. Ann Intern Med 1997; 127:796–803.

957

31. Cohen DA, Nsuami M, Etame RB, et al. A school-based Chlamydia control program using DNA amplification technology. Pediatrics 1998; 101:E1. 32. Paavonen J. Genital Chlamydia trachomatis infections in the female. J Infect 1992; 25(Suppl 1):39–45. 33. Bevan CD, Johal BJ, Mumtaz G, Ridgway GL, Siddle NC. Clinical, laparoscopic and microbiological findings in acute salpingitis: report on a United Kingdom cohort. Br J Obstet Gynaecol 1995; 102:407–14. 34. Eschenbach DA, Wo¨lner-Hanssen P, Hawes SE, Pavletic A, Paavonen J, Holmes KK. Acute pelvic inflammatory disease: associations of clinical and laboratory findings with laparoscopic findings. Obstet Gynecol 1997; 89:184–92.