## Abstract

A total of 107 sexually active women, aged 18–35 years, was recruited through family planning clinics in four centres in different countries. Each woman underwent two or three gentle but thorough and systematic vaginal inspections using a consistent technique with colposcopic magnification over a 4–6 month period to look for changes in vaginal and cervical appearance which might be related to sexual intercourse, tampon use, contraceptive method used, cigarette smoking or other environmental factors. Obvious changes in appearance were photographed at ×10 magnification. These appearances' or conditions' were classified according to a modification of the recommendations of a workshop sponsored by the World Health Organization, the Population Council and the Conrad Program. Most of these alterations in the appearance of the vaginal epithelium were judged to be of such minor clinical importance that they have been termed conditions' or changes in appearance' rather than lesions'. In all, 56 conditions' or appearances' were noted in 314 inspections, the commonest being petechiae (53.6%). Potentially significant conditions justifying the term lesions' (three microulcerations, two ecchymoses, five abrasions and one mucosal tear; 3.5% of inspections) usually healed spontaneously and disappeared rapidly. The incidence of these conditions was highest when the inspections followed intercourse in the previous 24 h (25.2 versus 14.2%; P < 0.0008), or tampon use (32.4 versus 15.9%; P < 0.0001). These processes may be regarded as a reflection of regular minor trauma to the vaginal epithelium.

## Introduction

The vagina is an organ which is regularly visualized, but rarely closely inspected, especially those parts of the fornices which may be obscured by the cervix or a vaginal speculum during routine gynaecological examination. The posterior fornix may, in fact, be the area of vaginal surface which is most likely to experience minor trauma from sexual intercourse, tampon use, diaphragm or vaginal ring use, or manipulation of a speculum. Surprisingly few data exist describing the appearance of the vaginal epithelium in response to environmental factors. In the course of normal activity, the vagina may be subject to regular and sometimes substantial trauma. The effects on the vaginal epithelium of sexual intercourse, tampon use, barrier methods of contraception, and childbirth have not been well documented.

A number of vaginal ring delivery systems are being developed for contraception (Shoupe and Mishell, 1992; World Health Organization, 1990) and menopausal hormone replacement therapy (HRT) (Ayton et al., 1996). These hormone-releasing rings have great promise, since they provide constant hormone concentrations, can be controlled by the woman herself and have been well tolerated by most users (Weisberg et al., 1995). However, little is known about the local effects of intravaginal devices upon the vagina. Data from women using large and relatively inflexible ring pessaries for genital prolapse have shown that they may occasionally cause ulcers but that the vagina has good healing properties (Nicholls, 1993). The possibility of an occasional adverse effect of intravaginal contraceptive rings on the integrity and health of the vaginal epithelium has been suggested by a recent, uncontrolled study of women using a levonorgestrel-releasing contraceptive vaginal ring (CVR) (Bounds et al., 1993). Since genital ulceration is a factor in the transmission of sexually transmitted diseases (STD) including the human immunodeficiency virus (Plummer et al., 1991), it is vital that it be known whether specific intravaginal devices affect the integrity of the vaginal epithelium.

Since little is known about the appearance of the vaginal epithelium in women who are not using the CVR, this study was designed to gather information on the role of factors such as sexual intercourse, tampon use, cigarette smoking and conventional contraceptive use in producing changes in the vaginal epithelial surface. This information will provide a basis for the assessment of factors which may alter the vaginal epithelium, including local effects of a CVR. The study also aimed to standardize the methodology of precise vaginal inspection to assess the incidence of minor changes in surface appearance.

## Materials and methods

This study was conducted in four centres in Helsinki, Finland (centre 1), Los Angeles, USA (centre 2), Santo Domingo, Dominican Republic (centre 3) and Sydney, Australia (centre 4) under the sponsorship of The Population Council, New York. A total of 107 sexually active women aged between 18 and 35 years with regular menstrual cycles (28 ± 7 days) was enrolled. Volunteers were approached through regular clinics or by advertisement within local family planning networks and were recruited over a 2 month period. The study was approved by local ethics committees and the Institutional Review Board of The Population Council. Subjects were given a detailed information sheet, and gave written informed consent before participation.

In order to obtain information about a wide variety of factors which might affect the vaginal epithelium, the women enrolled were sexually active and users of a range of different contraceptive methods, and some were regular tampon users. Women with obvious cervical ectopy, cervicitis, human papilloma virus, vaginal discharge or possible infection or other cervical or vaginal lesions were excluded from the study. A decision was made not to study bacterial vaginal flora because of the complexities of uniform and systematic cultures and identification between four centres in four different countries. Vaginal flora was only studied if there was symptomatic or obvious vaginal discharge which occurred during the course of the study. Participants were required to have had a normal Pap smear in the preceding 6 months.

Detailed demographic data were obtained which included subject age, contraceptive use, smoking habits, symptoms of vaginitis, use of a vaginal douche, frequency of sexual intercourse in the 2 weeks prior to examination and date of the last episode of sexual intercourse. The women were also asked to provide details of their menstrual cycle and of tampon use and timing, including type of tampon used. Tampons were not used by any of the subjects in Santo Domingo. At the initial screening, the subjects were given a full explanation of the study and their informed written consent was obtained. Speculum and bimanual examinations were performed. Each woman was assessed on three occasions over a 4–6 month period.

All clinics used the identical colposcopy equipment and one of the investigators (M.L.) visited each site, to ensure uniformity of observation techniques and descriptive terminology. Findings were reviewed in two meetings of investigators and photographs of lesions from all clinics were reviewed at the Population Council. Although many of the conditions observed in this study were of a size and severity that would not ordinarily be described as a lesion', the term lesion' has, for convenience, been used in some places in this report regardless of clinical importance of the observation. A standard recording form was used throughout. Initial inspection used ×6 magnification, which was increased to ×10 or greater to assess any lesion. After gentle irrigation with normal saline, the cervix and fornices were systematically inspected. Manipulation of the speculum was done as gently as possible. Acetic acid 3 or 5% was applied and suspicious aceto-white areas biopsied (these were all on the cervix). All lesions or changes in surface appearance seen with the naked eye or on colposcopy were recorded on diagrams indicating the region of the fornices, vaginal walls or vulva where they occurred and the nature of the lesion was noted (World Health Organization Global Programme on AIDS, 1996). Emphasis was placed on the detection of all changes in appearance, no matter how minor. Colour photographs were taken (at a magnification of ×10 or greater) of any atypical appearances using an identical Nikon camera in each centre. Cervical lesions were noted and recorded, although they were not the main focus and have not been described in this paper. Women with cervical intraepithelial neoplastic lesions at the first visit were excluded from the study. Vaginal bacteriological cultures were only carried out if clinically indicated.

Atypical appearances of the vaginal epithelium were classified according to a modification of criteria recommended by a Workshop sponsored by the World Health Organization, the Population Council and the Conrad Program (Table I) (World Health Organization Global Programme on AIDS, 1996). With a potentially significant lesion, the woman was asked to return for a repeat inspection 4–10 days later.

### Behavioural and demographic data which could influence vaginal appearance

One hundred and seven volunteers were aged between 18 and 35 years. Twenty-six women currently smoked cigarettes and 81 did not. All were sexually active at enrolment but two ceased sexual activity during the course of the study. Current contraceptive use included 28 women using the combined pill with or without condoms, 28 using condoms alone, 18 with male or female sterilization, 14 using an intrauterine device or system, eight using female barriers, five using Norplant and six using no contraception. Information was sought about frequency of vaginal intercourse in the 2 weeks and the 24 h before each colposcopic inspection. On 103 occasions, vaginal intercourse occurred during the 24 h prior to inspection.

Fifty-three of 107 women were regular tampon users (none in Santo Domingo). Most (38 women, 72%) used regular sized tampons only and 15 women (28%) a mixture of regular or super', dependent on personal choice and on the heaviness of their daily flow. None used deodorized tampons, and none gave a history of douching. Fifty women (47%) experienced occasional symptoms of vaginitis, three frequently' and 54 never'. Thirty-one women complained of recent vulvo-vaginal itch, 22 of vaginal discharge and two of odour, but none had an active vaginitis at the time of inspection. In the month before vaginal inspection, 13 women had taken broad spectrum antibiotics. Sixteen women had used vaginal antifungal agents in the previous month, and one had inserted yoghurt into her vagina. Twenty of the women sometimes experienced vaginal dryness during intercourse.

## Results

Many of the minor changes were difficult to see with the naked eye (over 80%), and were only seen clearly with ×6–10 magnification with the colposcope. Some of these appearances (26% of the petechiae) were only seen following application of acetic acid, and approximately one-half were classified as few' (fewer than five petechiae) and/or small' (<0.5 cm diameter in any direction). Frequency of reporting of petechiae decreased from the beginning to the end of the study and may have been related to minor changes in technique of manipulation of the speculum. In the Helsinki centre, petechiae were noted in nine of the first series of inspections (36.0%), four of the second inspections (17.4%) and only one of the third inspections (4.3%). Those thought to be of potential clinical significance' [including all microulcerations, epithelial tears, and the larger ecchymoses and abrasions (>0.5 cm in any direction)] were the three microulcerations, two ecchymoses, five abrasions and one epithelial tear. One microulceration was probably herpetic and the other two (of similar appearance) may also have been of similar aetiology, but culture did not confirm these. Three abrasions followed recent intense sexual intercourse, one followed tampon use and the other one had no obvious precipitating factors. The one epithelial tear' was a small laceration' of the superficial epithelium probably related to manipulation of the speculum.

There were no statistical differences in the incidence, location or nature of the atypical appearances observed in relation to contraceptive use. For example, of the 84 inspections in women who used condoms 18 (21.4%) had minor vaginal changes (10 petechiae, five erythema, ecchymoses or abrasions, one oedema; one aceto-white patch; one microulceration). Amongst users of other types of contraception, the incidence of atypical vaginal appearances ranged from 13.1 to 26.7%.

In Sydney, all of the women with atypical appearances were tampon users, and in Helsinki, most of them were. In these two centres, there was evidence that current tampon use made the detection of minor changes (mainly petechiae and erythema substantially more likely. An atypical appearance was noted on 12 out of 37 occasions (32.4%) when a tampon had to be removed for the inspection, compared with 44 out of 277 inspections (15.9%) when a tampon was not removed just before inspection (χ2 = 44.27; P < 0.0001). Residual lint from a tampon sometimes remained in the vaginal fornices. No microulcerations were noted in relation to tampon use.

A history of vaginitis, increased discharge, odour, itch, vaginal dryness or recent antibiotic use did not appear to be associated with the presence or absence of vaginal conditions or lesions'.

## Discussion

This study has confirmed the suspicion that minor superficial vaginal conditions, which do not merit the pathological implication of the term lesion' are fairly common in women under normal circumstances (seen in 56 out of 314 inspections: 17.8%; and in 41 of 107 women: 38.3%). The commonest conditions were petechiae (usually minor), and these seemed to disappear quickly. A previous study failed to show any traumatic effects on the vaginal epithelium in diaphragm users, but did not use magnification with the colposcope for inspection (Soper et al., 1991). These authors demonstrated a change in vaginal flora in diaphragm users. The ecology of the vaginal flora is recognized to be susceptible to numerous endogenous and exogenous influences, but a recent study (Davies et al., 1992) has failed to show any change in vaginal flora in users of a combination CVR.

Petechiae and erythema were the most commonly noted changes in epithelial appearance and were most likely to be associated with minor trauma to the epithelium, including recent sexual intercourse, recent tampon use and sometimes speculum manipulation. These changes in appearance could be regarded as normal minor trauma to the vaginal epithelium, and usually healed very rapidly. A minority of the atypical appearances were only clearly visible following acetic acid application, and acetic acid may have had a promoting' effect on an already mildly traumatized epithelium. Small variations in the observation of erythema, petechiae and abrasions were noted between centres in spite of concerted efforts at standardization, and it has to be emphasized that the great majority of the conditions' or even lesions', noted in this study would have been overlooked or disregarded in the clinical situation because of their minor' or benign' appearance.

Few of the demographic factors considered in this study appeared to be associated with a significant increase in atypical appearances of the vaginal epithelium, but the numbers of subjects were too small for a definitive conclusion to be reached.

A highly significant relationship of colposcopically detected minor changes to current tampon use was found in this study, although no serious lesions were noted. Several previous reports have emphasized the possibility of potentially serious lesions, such as microulcerations, during prolonged use or in tampon use between menses (Barrett et al., 1977; Friedrich and Siegesmund, 1980; Jimmerson and Becker, 1980). There is evidence to suggest that microulcerations were most likely when superabsorbent tampons (no longer marketed) were worn at times other than during menstruation (Friedrich and Siegesmund, 1980).

There was also a significant positive association between the number of minor vaginal conditions and smoking (28.6% compared with 14.3; P = 0.0001). Cigarette smoking has a known detrimental effect on squamous epithelium (Szarewski et al., 1996) and constituents of smoke concentrate in cervical epithelium. It is possible that smoking may make the epithelium and superficial vasculature more susceptible to minor changes (Szarewski et al., 1996).

A highly significant association was seen between the observation of epithelial conditions in women experiencing sexual intercourse in the previous 24 h (25.2%) compared with those with longer intervals (14.2%; P < 0.0008). A positive association was also noted between high frequency of intercourse in the previous 2 weeks and the presence of a variety of atypical appearances. Erythema of the vaginal introitus and lower vagina, abrasions, ecchymoses and prominent petechial lesions of the vaginal fornices were seen on several occasions following recent intercourse, and infrequently at other times. These findings are in agreement with a small study where women were inspected with the colposcope within 6 h after intercourse and found to have telangiectasia, broken capillaries and microabrasions' in 61.1% of inspections compared with 11.1% of cases after abstinence (Norvell et al., 1984). It is not surprising that minor degrees of surface trauma are frequently seen in the vagina, since major trauma and epithelial tears also occur at times following voluntary sexual intercourse (Sill, 1987).

In a vaginal inspection study in long-term levonorgestrel-releasing contraceptive vaginal ring users, the changes in surface appearance seen in 48/149 subjects (34%) were mainly erythematous patches, but a small number of chronic ulcerations were seen (Bounds et al., 1993). These ulcerations were usually noted first as erythematous areas in the vaginal fornices, and were sometimes aceto-white on staining. Biopsy of these obvious chronic ulcerations confirmed the presence of a chronic inflammatory cell infiltrate, often with focal spongiosis, focal thinning or patchy absence of the epithelium and dilated stromal vessels, and occasionally with chronic granulation tissue. Ulcerations were generally only detected after more than 6 months of continuous use of the levonorgestrel-releasing ring, although erythematous patches were noted much earlier. Most lesions disappeared fairly rapidly following removal of the ring, but a few persisted for many months. Changes in vaginal flora were not investigated. There remains a need for careful, controlled assessment of the local influence of contraceptives in the vagina and for an increased understanding of the influence of non-contraceptive devices and sexual behaviours on vaginal surface appearance and microbiology function.

Several studies have demonstrated that frequent application of nonoxynol-9 (N-9) may be related to epithelial disruption of the cervix and vagina (Niruthisard et al., 1991; Roddy et al., 1993) and genital ulcers (Kreiss et al., 1992), although not all vaginal delivery systems for nonoxynol-9 appear to be associated with these findings (Poindexter et al., 1996). The cervical epithelium appears to be less resilient to spermicides than the vulvo-vaginal epithelium (Poindexter et al., 1996). The lesions noted on the ectocervix included `redness, white epithelium, ulceration, mosaic petechiae and squamous metaplasia with redness being the most common occurrence'. The majority of lesions found in the N-9 studies are of a more serious type than those noted in the present study, which makes it unlikely that the results of the N-9 studies are inflated by reports of minor atypical appearances occurring as a result of normal wear and tear on the vagina. Other potentially cytotoxic substances such as hydrogen peroxide are used in some communities as vaginal contraceptive agents, and could be expected to produce more pronounced local lesions than seen in the present study if use is repeated over some years (Strohmer et al., 1997).

This study has demonstrated that a number of minor conditions can be observed on the vaginal epithelial surface of normal women provided that care is taken with inspection and that colposcopic magnification is used. Some of these changes only became visible following application of acetic acid. Minor trauma appears to be a precipitating factor with petechiae, abrasions and erythema being noted more frequently in women who have recently had sexual intercourse or recently used a tampon. Speculum manipulation may also produce these appearances.

Table I.

Characteristics of vaginal epithelial lesions and other surface changes as assessed using the colposcope

Microulcer Abrasion Ecchymosis Petechial haemorrhage Sub-epithelial haemorrhage and swelling Erythema Oedema Epithelial tear
Epithelium Disrupted Partially disrupted Intact Intact Intact + swelling Intact Intact + swelling Disrupted
Blood vessels Intact or disrupted Intact or disrupted Disrupted Disrupted Disrupted Intact Intact Disrupted
Demarcation Sharp Diffuse Variable Sharp Variable Variable Variable Sharp
Peripheral reaction Variable Variable Variable Variable Variable No Variable Little or none
Slough Yes No No No No No No None
Size Any Any >0.5 cm each < 0.5 cm Any Any Any Any
Modified from World Health Organization (1966).
Microulcer Abrasion Ecchymosis Petechial haemorrhage Sub-epithelial haemorrhage and swelling Erythema Oedema Epithelial tear
Epithelium Disrupted Partially disrupted Intact Intact Intact + swelling Intact Intact + swelling Disrupted
Blood vessels Intact or disrupted Intact or disrupted Disrupted Disrupted Disrupted Intact Intact Disrupted
Demarcation Sharp Diffuse Variable Sharp Variable Variable Variable Sharp
Peripheral reaction Variable Variable Variable Variable Variable No Variable Little or none
Slough Yes No No No No No No None
Size Any Any >0.5 cm each < 0.5 cm Any Any Any Any
Modified from World Health Organization (1966).
Table II.

Characteristics of vaginal surface conditions observed during 314 inspections by colposcope in 107 women

Nature of lesion Number of observations by centre
Total (% of inspections)
Oedema  4 (1.3)
Erythema  9 (2.9)
Petechiae 14 10  30 (9.6)
Ecchymosis  2 (0.6)
Abrasions  5 (1.6)
Lacerations  1 (0.3)
Haemorrhage
Microulceration  3 (0.9)
Aceto-white  2 (0.6)
Total conditions 19 16 11 10  56 (17.8)
Total inspections 71 75 95 73 314
Nature of lesion Number of observations by centre
Total (% of inspections)
Oedema  4 (1.3)
Erythema  9 (2.9)
Petechiae 14 10  30 (9.6)
Ecchymosis  2 (0.6)
Abrasions  5 (1.6)
Lacerations  1 (0.3)
Haemorrhage
Microulceration  3 (0.9)
Aceto-white  2 (0.6)
Total conditions 19 16 11 10  56 (17.8)
Total inspections 71 75 95 73 314
Table III.

Incidence of vaginal conditions in women inspected within 24 h of vaginal intercourse

Presence of vaginal conditions Intercourse within less than 24 h of inspection
Yes No
Yes  26  30
No  77 181
Total inspections 103 211
Total conditions within 24 h of intercourse = 26 out of 103 inspections (25.2%).
Total conditions unrelated to timing of intercourse = 30 out of 211 inspections (14.2%).
Presence of vaginal conditions Intercourse within less than 24 h of inspection
Yes No
Yes  26  30
No  77 181
Total inspections 103 211
Total conditions within 24 h of intercourse = 26 out of 103 inspections (25.2%).
Total conditions unrelated to timing of intercourse = 30 out of 211 inspections (14.2%).
Table IV.

Appearance of 22 vaginal conditions seen at 77 inspections in women who smoked cigarettes

Type of lesion Numbers Location
Oedema Generalized
Erythema Introitus & lower third of vagina
Abrasion Lower third of vagina
Petechiae 13 Fornices, middle & lower third of vagina
Microulceration Anterior fornix
Type of lesion Numbers Location
Oedema Generalized
Erythema Introitus & lower third of vagina
Abrasion Lower third of vagina
Petechiae 13 Fornices, middle & lower third of vagina
Microulceration Anterior fornix
7
To whom correspondence should be addressed at: Department of Obstetrics and Gynaecology, University of Sydney, NSW 2006, Australia

This study was funded by the United Nations Population Fund and the Population Council, New York.

## References

Ayton, R.A., Darling, G.M., Murkies, A. et al. (
1996
) A comparative study of safety and efficacy of continuous low-dose oestradiol released from an intravaginal ring versus conjugated equine oestrogen vaginal cream in the treatment of postmenopausal urogenital atrophy.
Br. J. Obstet Gynaecol.
,
103
,
351
–358.
Barrett, K.F., Bledsoe, S., Greer, B.E. et al. (
1977
) Tampon-induced vaginal and cervical ulceration.
Am. J. Obstet. Gynecol.
,
127
,
332
–333.
Bounds, W., Szarewski. A., Lowe, D. et al. (
1993
) Preliminary report of unexpected local reactions to a progestogen-releasing contraceptive vaginal ring.
Eur. J. Obstet. Gynecol. Reprod. Biol.
,
48
,
123
–125.
Davies, G.C., Feng, L.X., Newton, J.R. et al. (
1992
) The effects of a combined vaginal ring releasing ethinyl-oestradiol and 3-ketodesogestrel on vaginal flora.
Contraception
,
45
,
511
–518.
Friedrich, E.G. and Siegesmund, K.A. (
1980
) Tampon-associated vaginal ulcerations.
Obstet. Gynecol.
,
55
,
149
–156.
Jimmerson, S.D. and Becker, J.D. (
1980
) Vaginal ulcers associated with tampon usage.
Obstet. Gynecol.
,
56
,
97
–99.
Kreiss, J., Ngugi, E. and Holmes K. (
1992
) Efficacy of nonoxynol-9 contraceptive sponge use in preventing heterosexual acquisition of HIV in Nairobi prostitutes.
JAMA
,
268
,
477
–482.
Nicholls, J. (
1993
) An unusual vaginal foreign body.
Aust. N. Z. J. Obstet. Gynaecol.
,
33
,
101
–102.
Niruthisard, S., Roddy, R.E. and Chutivongse, S. (
1991
) The effects of frequent nonoxynol-9 use on the vaginal and cervical mucosa.
Sex. Transm. Dis.
,
18
,
176
–179.
Norvell, M.K., Benrubi,G.I. and Thompson, R.J. (
1984
) Investigation of microtrauma after sexual intercourse.
J. Reprod. Med.
,
29
,
269
–271.
Plummer, F.A., Simonsen, J.N., Cameron, D.W. et al. (
1991
) Cofactors in male-female sexual transmission of human immunodeficiency virus type 1.
J. Infect. Dis.
,
163
,
233
–239.
Poindexter, A.N., Levine, H., Sangi-Haghpeykar, H. et al. (
1996
) Comparison of spermicides on vulvar, vaginal and cervical mucosa.
Contraception
,
53
,
147
–153.
Roddy, R.E., Cordero, M., Cordero, C. et al. (
1993
) A dosing study of nonoxynol-9 and genital irritation.
Int. J. STD AIDS
,
4
,
165
–170.
Shoupe, D. and Mishell D.R. Jr (1992) Contraceptive vaginal rings: efficacy and acceptability. In Sitruk-Ware, R. and Bardin, C.W. (eds), Contraception: Newer Pharmacological Agents, Devices and Delivery Systems. Marcel Dekker, New York, pp. 71–89.
Sill, P.R. (
1987
). Non-obstetric female genital tract trauma in Port Moresby, Papua New Guinea.
Aust. N.Z. J. Obstet. Gynaecol.
,
27
,
164
–165.
Soper, D.E., Brockwell, N.J. and Dalton, H.P. (
1991
) Evaluation of the effects of condom on the female lower genital tract.
Contraception
,
44
,
21
–29.
Strohmer, H., Ozturk, D. and Egarter, C. (
1997
) Use of hydrogen peroxide for vaginal contraception.
Hum. Reprod.
,
12
,
1599
.
Szarewski, A., Jarvis, M.J., Sasieni, P. et al. (
1966
) Effect of smoking cessation on cervical lesion size.
Lancet
,
347
,
941
–943.
Weisberg, E., Fraser, I.S., Mishell, D.R. Jr et al. (
1995
) The acceptability of a combined oestrogen–progestogen contraceptive vaginal ring.
Contraception
,
51
,
39
–44.
World Health Organization Task Force on Long-Acting Systemic Agents for Fertility Regulation (
1990
) Microdose intravaginal levonorgestrel contraception: a multicentred clinical trial. I. Contraceptive efficacy and side-effects
. Contraception
,
41
,
105
–124.
World Health Organization Global Programme on AIDS (1996) Manual for the standardization of colposcopy for the evaluation of vaginally administered products. WHO, Geneva , p. 15.