This supplement contains the report of a meeting convened by CONRAD and the World Health Organization (WHO) entitled ‘Assessing Inflammation and Epithelial Integrity in Vaginal Product Research’. It took place on November 19–21, 2003, in the Dominican Republic. The meeting addressed ways of detecting the adverse effects of vaginal products intended to reduce the risk of acquisition of HIV and sexually transmitted infections (STIs), before products move into Phase III effectiveness trials and are considered for approval for use.
Despite advances in the care of persons infected with HIV, the AIDS epidemic remains one of the most significant and daunting public health crises of recent times. Its effects are felt disproportionately by those areas with the fewest resources: 95% of new infections occur in the world's poorest countries, those least able to provide medical treatment and supportive care.1 Heterosexual (vaginal) transmission is responsible for most HIV infections in these countries.
Efforts at prevention are still focused mainly on behavior change, as no effective vaccine currently exists. Vaginal products that could be used topically before intercourse to prevent the acquisition of HIV and other STIs are receiving increased attention and support. Such products, referred to as ‘microbicides’, are being studied by public and private organizations and hold promise as being products that could be under control of the woman and offer protection against HIV/STIs.
DEMONSTRATING MICROBICIDE SAFETY AND EFFECTIVENESS
Before such products can be approved for use by women at risk of HIV/STIs, their safety and effectiveness must be demonstrated. For most new therapies, safety is the absence of ‘side effects’ that are usually unrelated to efficacy. In the case of microbicides, however, there are reasons to believe that the possible adverse effects of such products can not only cause discomfort, but also reduce efficacy.
This sequence of events is shown in the schematic (Figure 1). The central feature is ‘mucosal changes’, which could include breaks in the epithelium, inflammation, immune dysfunction, or effects that have not been identified (possibly changes in microflora). These changes could adversely affect the safety of the product by causing either discomfort only, or by creating portals of entry for pathogens or activation and recruitment of immune cells, thereby reducing effectiveness and increasing the chance of acquisition of HIV and other STIs.
THE NONOXYNOL-9 EXAMPLE
The vaginal spermicide nonoxynol-9 (N-9) has shown anti-HIV effects in laboratory testing, and was a promising candidate as a microbicide. In a clinical trial in which a gel containing 52.5 mg N-9 per dose was compared with an inactive (placebo) control, the use of more than 3.5 applications per day was associated with an increase rather than the hoped-for decrease in HIV acquisition.2 Breaches of the vaginal epithelium noted on clinical examination were associated with seroconversion, and it has since been speculated that the virucidal effect of N-9 was overcome by its disruption of the epithelial barrier and possibly also its induction of an increased number of target cells susceptible to infection.
EFFORTS TO DETECT EPITHELIAL CHANGES
Efforts to detect the adverse effects of vaginal products visually have been underway for a decade. In 1993, WHO convened a conference of experts to standardize the use of colposcopy, a procedure more often used to magnify the cervical epithelium as a diagnostic test in women with abnormal Papanicolaou smears. The result of that conference was a manual3 outlining a modified colposcopy procedure and terminology to be used in the assessment of the vaginal and cervical reaction to vaginal products.
This first WHO manual was revised in conjunction with a second expert review convened by CONRAD and WHO in 1999. The new manual, referred to as ‘Update 2000’, shortened the procedure by eliminating the use of the green filter and acetic acid, and simplified the recording of findings by limiting recording to the intactness of the epithelium and blood vessels, and eliminating the use of descriptors such as erythema, abrasion, etc.4
The manuscript by Njoroge et al.,5 in this supplement addresses how well the Update 2000 procedure has worked in the field. The CONRAD colposcopy study described in the paper by Ballagh6 has shed light on the value of colposcopy over naked eye examination, the amount of variation between observers, and two different methods of magnified observation. The development of a photo atlas for clinical examination in microbicide studies described by Bollen et al.7 has helped further standardize the documentation of findings.
In recent years, the benefits and limitations of colposcopy have become clearer. Despite efforts at standardization, the decision as to what to record remains somewhat subjective. Standard colposcopy requires reliable electricity, expensive equipment, and specific training, all of which increase the cost and complexity of vaginal product trials. There is interest in developing a less cumbersome method of magnified visual assessment. Hand-held colposcopes have been studied and are described in the manuscripts by Sellors et al.8 and Walmer et al.9 The adaptation of digital photography, described by Jamshidi and Blumenthal,10 is especially promising.
A major development in assessing microbicide safety is the recognition that not all important epithelial changes will be detectable by visual means. It has become increasingly apparent that the use of vaginal products can lead to the activation of immune cells, with the production of proinflammatory cytokines, which cause the recruitment and activation of immune cells that are targets for HIV infection. Rabbit studies have shown flattening and disruption of the epithelial lining, increased leukocyte infiltration, and general signs of inflammation such as edema and vascular congestion, together with the release of proinflammatory cytokines that may be quantified in cervicovaginal lavage (CVL) using a refined rabbit vaginal irritation model.11 Proinflammatory cytokines were also increased in women who applied N-9. When these CVLs were incubated with HIV latently infected cells, there was an induction of HIV replication, which reached a maximum with CVLs collected 60 h after N-9 administration, showing one of the possible links between mucosal inflammation and an increased risk of HIV transmission.
The relative contribution of visually detectable epithelial changes and inflammation to an increased risk of HIV is not known. It is clear that epithelial disruption and inflammation can develop independently, although excessive inflammation may cause epithelial disruption and vice versa. It will be important for the protective effect of the product to last longer than any deleterious effect on the mucosa. Moreover, in a clinical trial it is impossible to determine the exact sequence of events or to identify the exact time of HIV infection. In the light of these unanswered questions, it was felt that another conference of experts was warranted, this time to address not just colposcopy but also other methods of detecting the adverse effects of vaginal products.
THE 2003 CONFERENCE
The 2003 conference was attended by 68 participants from 13 countries, and included representatives from CONRAD and WHO as well as the US National Institutes of Health, the US Centers for Disease Control and Prevention, the Alliance for Microbicide Development, the International Partnership for Microbicides, Family Health International, Johns Hopkins University, and institutions in Benin, Brazil, the Dominican Republic, India, Kenya, Nigeria, South Africa, Switzerland, Thailand, Uganda, the United Kingdom, the United States of America, and Zimbabwe. Plenary presentations addressed colposcopy, other visual means of assessment, and non-visual means of assessment, and are represented by the articles in this supplement. After the plenary presentations, the participants were divided into four break-out groups, and each one was given a set of questions to be addressed. Their answers, which constitute recommendations for moving forward in the field of assessing microbicide safety, appear below.
Break-out Group 1: Review the methods of colposcopy, identify areas in the manual in need of updating, and identify key research needs
Members of Break-out Group 1
Francisco Alvarez, Eliana Amaral, Susan Ballagh, Liesbeth Bollen, Vivian Brache (rapporteur), Elizabeth Bukusi (chair), Mike Chirenje, Leila Cochón, Mitchell Creinin, Fernand Aime Guedou, Kamal Hazari, Smita Joshi, Antonia Kwiecien, Maya Lulla, Cecy Manjila, Christine Mauck, Florence Mirembe, Jocelyn Moyes, Nelly Mugo, Clemensia Nakabito, Betty Njoroge, Lawrence Odusoga, and Bina Pandey.
Questions addressed by Break-out Group 1
- Is an updated manual needed?
- If so, what are the changes needed, and how can such an update best be distributed?
- Is additional research needed to refine the technique of colposcopy and if so, what should it address?
Recommendations from Break-out Group 1
An updated manual was felt to be needed and has been written. Copies are available as printed manuals and on CD from CONRAD (1-703-524-4744) and on the CONRAD and WHO websites (www.conrad.org and www.who.int/reproductive-health). The main revisions were:
- The reintroduction of descriptors such as erythema, edema, abrasion, etc. Although it was still felt that the intactness of the blood vessels and epithelium were the key variables to analyse, it was found that use of clinical descriptors facilitated communication among investigators and allowed for the more accurate recording of findings. With clear definitions of what constitutes the various clinical descriptors, the intactness of the blood vessels and epithelium can be inferred.
- Clarification of recording of the severity of findings and worsening of findings.
- Standardized recording of ectopy and bleeding (recognizing their possible importance as factors that may be related to HIV acquisition).
- Clarification of recording normal and non-colposcopic findings.
- Redesigning of the Data Collection Form to allow the recording of multiple findings from multiple time points on one sheet.
In addition, certain research questions were outlined, including:
- How long after a Papanicolaou smear should colposcopy be performed to minimize the effects of the smear on colposcopy results?
- What is the relationship between cervical ectopy, inflammation, and STI/HIV acquisition?
- What are the normal changes in vascular patterns that can be expected in women with spontaneous menstrual cycles and women using exogenous hormones? For example, does thinning of the epithelium seen in the luteal phase and in women using certain contraceptive methods make vessels appear more prominent? If so, it could be important to differentiate this from a change in vascular prominence observed in a trial of a new vaginal product that could indicate the presence of an inflammatory infiltrate or increased flow to the cervix, possibly increasing the risk of HIV.
- How is erythema best assessed? (This was also raised by Group 2.) On the one hand, it should be technically possible to make the redness of a recorded image match the redness actually observed. It should also be technically possible to define certain degrees of redness as erythema. However, erythema in one woman (who is, for example, anemic at baseline but experiences irritation from a product) may be normal coloration for another woman (not anemic and not experiencing irritation).
- What is the natural history of colposcopic findings? This is clearly related to the task of Group 4 to determine what effects of products indicate an increased risk.
Break-out Group 2: Review new developments in techniques of visualization other than colposcopy and identify key research needs
Members of Break-out Group 2
Teresa Abercrombie, Ahymala Balasubramanian, Lynn Bradley, Hugo De Vuyst (chair), Miguel Espinoza, Roxanne Jamshidi, George Kovalesky (rapporteur), Alfred Poindexter, Angela Rivers, John Sellors, and Dave Walmer.
Questions addressed by Break-out Group 2
- What characteristics should a visual assessment technique have in order to be useful in developed and developing country settings?
- Of the techniques presented, do any warrant additional research, and if so, what should it address?
- Alternatively, should we design something new? If so, what characteristics should it have to function optimally in developed and/or developing country settings? How should it be studied?
Recommendations from Break-out Group 2
- The same technique should be used in developed and developing countries. This may force some compromises on the technique, but it is important that the technique be standardized everywhere.
- The ‘best’ technique will depend on the information that is being sought (addressed by Group 4).
- Current technology can be adapted; there is no need to start over. A table was constructed (Table 1) comparing attributes of the naked eye examination, a hand-held colposcope (AviScope®), digital photography, and colposcopy. Colposcopy clearly provided the best quality of examination, but was the most difficult and costly to use. The AviScope® is undergoing improve ment, which should improve image quality. Digital assessment (so-called digital assessment of the reproductive tract, or DART) shows promise both in producing quality images and in being easy and inexpensive to use.
A list of research questions included:
- Can the light source wavelength and intensity be standardized, because these affect the perception of color?
- How much magnification is ‘enough’? The same question holds for resolution.
- How important is depth perception, i.e. does a binocular device have significant advantages over a monocular device?
Break-out Group 3: Review new developments in non-visual assessment (biopsy techniques, inflammatory markers, and tests for occult blood) and identify key research needs
Members of Break-out Group 3
Judith Absalon, Kurt Barnhart, Susan Cu-Uvin (chair), Teresa De Pinares, Gustavo Doncel, Raina Fichorova, Polly Harrison, Joann Kuruc, Kim Linton, Naomi Low-Beer (rapporteur), Joanne Luoto, Kenneth Mayer, Isaac Malonza, Thomas Moench, Jeffrey Peipert, Jill Schwartz, and Debra Wiener.
Questions addressed by Break-out Group 3
- What are the relative strengths and weaknesses of each of the techniques presented?
- Are there other non-visual assessment techniques that could be helpful in identifying adverse effects of vaginal products?
- How should research proceed for non-visual techniques?
Recommendations from Break-out Group 3
- The relative strengths and weaknesses of each of the non-visual techniques presented can be described as follows:
- Biopsy is the only method that can provide direct information on the inflammatory status of human cervicovaginal mucosal tissue, including infiltrating CD4-positive HIV-1 target cells (see paper by Low-Beer).12 There are existing data from animal studies, and the technique has been used in clinical trials. However, there is a risk of sampling error (not taking the sample from an affected area), and even when apparently abnormal areas have been biopsied, there has been little correlation between visual changes and biopsy results. In addition, the procedure is invasive, there is a lack of normative data, and the biopsy itself may increase the risk of disease transmission. The group recommended that biopsy continue to be used in selected subgroups to validate and understand the meaning of colposcopic findings. Biopsies should be taken from both pre-determined sites (preferably including the stratified squamous and columnar epithelia and the transformation zone) and where abnormal colposcopic findings have been noted. In addition, extended animal studies in rabbits and mice should be carried out.
- Cervicovaginal lavagefor inflammatory biomarkers Biomarkers indicate changes at a cellular, biochemical, or genomic level that can provide quantitative data on normal or pathological processes. Inflammatory mediators and particularly cytokines are induced in the female genital tract by infectious and non-infectious inflammatory conditions. Proinflammatory cytokines play a critical role in HIV-1 pathogenesis.11,13
- Using CVLs for inflammatory markers is non-invasive, samples the entire cervico-vaginal tract, and can provide information about multiple parameters, e.g. cytokines, cells, blood, innate immunity, and the presence of STIs such as herpes simplex virus. Clinical studies have shown a correlation of proinflammatory cytokines detectable in CVLs with HIV-1 viral load and bacterial infection. On the other hand, the use of CVL cytokines to guide clinical microbicide trials has been delayed by the lack of normative values for the most relevant markers, and the limited information on the effect of factors such as genetics, sexual intercourse, exposure to semen, condom use, the menstrual cycle and other factors that may cause product independent inter-subject variability.13
- The group recommended that more preclinical research be done to define the most relevant markers by correlation with histopathology and increased HIV susceptibility. CVLs could be collected in a subset of participants in ongoing microbicide trials at selected sites and processed at a later date. The method of collection and storage would have to be standardized. It was suggested that 10 ml saline (0.9% NaCl) be applied and collected by the clinician, with centrifugation and separation of supernatant and pellet on site as quickly as possible. Hands-on training of involved personnel would be important. Samples could be collected at baseline, early in dosing, in the middle of dosing, and at the end, with consideration given to removing the product after the last dose and a final sampling occurring 3–4 days after the last dose.
- Tests for blood
- The detection of minute amounts of blood is described in the paper by Moench et al.14 This technique is non-invasive and may be a marker for breaches in the epithelium. It may also help with the interpretation of other markers and the source of HIV. However, determining the source of any blood detected would be difficult, inter-subject variability is expected to be high, there is a narrow window for testing to avoid menstrual bleeding, and there are no normative data.
- The group recommended that the technique be studied further to determine if it can be standardized and normative data obtained, and to assess possible interference by products (in-vitro studies).
- There may be other non-visual assessment techniques that could be helpful in identifying the adverse effects of vaginal products. The preclinical development of new markers of inflammation (e.g. nuclear factor kappa B), the validation of markers versus histopathology and immunoinflammatory characterization of mucosal infiltrates using animal models, and the correlation of markers with increased susceptibility to HIV infection, again using animal and in-vitro models, are crucial to determining the relevance of new markers. To that end, Group 3 suggested a stand-alone clinical study in healthy women not at risk of STIs to correlate colposcopy, histology, and inflammatory markers (possibly using N-9 as a positive control, a placebo, and a microbicide candidate).
- Other research could involve the staining of collagen to identify epithelial breaches. This would require animal studies. In addition, consideration should be given to assessing the impact of microbicides on the endometrium, assessing mucosal function (e.g. innate defences and permeability), and looking at cytokine genetic polymorphisms.
Break-out Group 4: Recommend future research to distinguish product-related effects from other factors that can affect the epithelium, and determine which epithelial changes indicate an unsafe product
Members of Break-out Group 4
Deborah Anderson (chair), Marianne Callahan, Anne Coletti, Tim Farley, Martha Jacob, Maryse Joseph, Pekka Lahteenmaki, Elizabeth Raymond (rapporteur), Lydia Soto-Torres, Lut Van Damme, and D'Nyce Williams.
Questions addressed by Break-out Group 4
- How can this type of information be gained from studies already planned?
- Is it possible to design feasible studies to look at these two questions? If so, how?
Recommendations from Break-out Group 4
Epithelial disruption and inflammation have been shown to follow vaginal product use, but may also result from other factors, such as mechanical trauma from intercourse itself, exposure to semen, the use of vaginal products other than the one being assessed, endogenous or exogenous hormonal factors, infections, and the very examination carried out to detect epithelial changes. The fundamental questions, therefore, include how to detect epithelial changes, which products cause changes, and, in particular, which changes and what intensity of changes are associated with an increased risk of STIs/HIV.
- Animal Model Research
- A model using macaques and simian immunodeficiency virus (SIV) was suggested. First, colposcopy findings would be compared with vaginal cytokine responses after treatment with N-9. The effects of dosage, number of exposures, and time course of effect could be explored. If the inflammatory cytokine response is similar to that which has been observed in women after multiple N-9 exposures, a challenge study could be conducted. This would involve an SIV vaginal challenge after product washout, during the subsequent inflammatory response. Such research could directly test the hypothesis that cervical/vaginal inflammation increases the presence of markers that predict susceptibility (as opposed to just looking for associations in clinical trials). This macaque model could be used to study the potential inflammatory effects of other microbicide products, and associated SIV susceptibility.
- Phase 1 and 2 Safety Trials Comparisons between colposcopic and cytokine data should continue in Phase 1 and 2 safety trials in low-risk women. Trials could then be extended to studies of high-risk women from HIV/STI endemic areas to test the hypothesis that pre-existing vaginal inflammatory conditions increase susceptibility to the proinflammatory effects of vaginal products. Women from STI endemic areas could be randomly assigned into (1) no treatment; (2) placebo; or (3) N-9 treatment groups. A baseline CVL and colposcopic examination would be carried out to determine the presence of a pre-existing inflammatory condition. Any identified curable STIs would be treated. The participant would administer her assigned product vaginally twice a day for 3 days while abstaining from intercourse. Colposcopy and cytokine profiles would be obtained at the end of the treatment period, and compared with baseline values. This procedure could be used to examine the effects of other candidate microbicide products by obtaining colposcopic findings and vaginal cytokine levels in women with and without inflammatory vaginal conditions.
- Phase 3 Clinical Trials Performing colposcopy on the large number of women necessary for Phase 3 trials is not feasible. If specimens that are simple to obtain (e.g. self-collected vaginal swabs) can be used to measure the markers associated with an increased risk of HIV-1 acquisition, they could be collected monthly and analysed to see if inflammatory markers actually do correlate with HIV acquisition.
The need for products with which women can protect themselves from HIV is undeniably urgent. The need to ensure that such products are both safe and effective is obviously a critical part of developing such products. However, little is known about normal cervicovaginal anatomy and physiology, especially the changes that occur in the presence of irritants and disease. Colposcopy as a means of visual assessment of such changes has been standardized and should continue to be used to detect breaches in the epithelium, while possible further refinements should be considered. Animal models should be developed to identify other markers of increased risk, and efforts at identifying such markers in women intensified. The correlation of visual changes and markers of inflammation with the risk of HIV/STIs is critical.
1. Lamptey PR. Reducing heterosexual transmission of HIV in poor countries. BMJ
2. Van Damme L, Ramjee G, Alary M, et al. COL-1492 Study Group. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomized controlled trial. Lancet
, 2002;360:971-977 [Erratum in: Lancet
3. World Health Organization. Manual for the standardization of colposcopy for the evaluation of vaginally administered products. Geneva: WHO; 1995.
4. World Health Organization/CONRAD. Manual for the standardization of colposcopy for the evaluation of vaginal products, Update 2000. Geneva: CONRAD/WHO; 2000.
5. Njoroge B, Chirenje ZM, Pandey B, et al., The CONRAD/WHO Update 2000 colposcopy manual: how well is it working? J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3):S150–S151.
6. Ballagh S. Factors affecting the reproducibility and variability of colposcopy for product development: review of current literature. J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3):S152–S155.
7. Bollen LJM, Kilmarx PH, Tappero JW. Interpretation of genital findings in microbicide safety trials: review of the ‘photo atlas for microbicide evaluation’. J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3):S156–S159.
8. Sellors JW, Winkler JL, Kreysar DF. Illumination, optics, and clinical performance of a hand-held magnified visual inspection device (AviScope™): a comparison with colposcopy. J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3):S160–S166.
9. Walmer DK, Littman E, Rodriguez G, et al. Portable colposcopy in low resource settings. J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3): S167–S170.
10. Jamshidi R, Blumenthal P. Digital assessment of the reproductive tract (DART). J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3):S171–S173.
11. Doncel GF, Chandra N, Fichorova RN. Preclinical assessment of the proinflammatory potential of microbicide candidates. J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3):S174–S180.
12. Low-Beer N. The role of biopsy in vaginal microbicide development. J Acquir Immune Defic Syndr
, 2004;37(Suppl. 3):S181–S183.
13. Fichorova RN. Guiding the vaginal microbicide trials with biomarkers of inflammation. J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3): S184–S193.
14. Moench T, Mehrazar K, Cone R, et al. Sensitive methods to detect epithelial disruption: tests for microhemorrhage in cervicovaginal lavages. J Acquir Immune Defic Syndr
, 2004; 37(Suppl. 3):S194–S200.