Probiotics enhance the clearance of human papillomavirus-related cervical lesions

a prospective controlled pilot study

Abstract

Probiotics have been proposed for a number of urogenital infectious conditions. In this study, we examine a possible effect on human papillomavirus (HPV)-related precancerous lesions in cervical cytology. We conducted a prospective controlled pilot study, in which 54 women with an HPV+low-grade squamous intraepithelial lesion diagnosis in their PAP smear were followed for 6 months. The intervention group consumed a daily probiotic drink during the study period; the control group received no treatment, according to common care policy. Outcome measures were the control PAP smear and HPV status after 6 months. Probiotic users had a twice as high chance of clearance of cytological abnormalities (60 vs. 31%, P=0.05). HPV was cleared in 19% of control patients versus 29% of probiotic users (P=0.41). This exploratory pilot study suggests that the probiotic studied promotes the clearance of HPV-related cytological abnormalities. If confirmed, this would represent an entirely new option to manage cervical cancer precursors.

Introduction

Three to six percent of PAP smears show abnormalities, most of which are minor cytological changes (^{Arbyn et al., 2009; Brismar-Wendel et al., 2009}). The majority of these minor changes regress without treatment, especially in women younger than 30 years of age. The widely recommended policy in the case of these abnormalities is to await spontaneous regression for a period of time before eventual surgical treatment. Low-grade squamous intraepithelial lesions (LSIL) are highly associated with high-risk human papillomavirus (HPV) infection, and the clearance of both shows a close temporal relationship (^{Syrjanen, 2007}). In Belgium, women with LSIL undergo a repeat cytology (PAP) test after 6 months to decide whether they should be referred to a colposcopy unit or be referred back to the screening interval of 3 years. Although small differences exist in screening, follow-up, and treatment strategies between countries, conservative management of LSIL is the standard of care in many industrialized countries. This 6-month period of watchful waiting offers the opportunity to study measures that can be taken to enhance the possibility of spontaneous clearance. To date, only smoking cessation has been offered as advice to patients.

Both oral and local probiotics have been examined and have shown some efficacy in the context of urogenital health (^{Spurbeck and Arvidson, 2011}): the best evidence exists for the prevention and treatment of bacterial vaginosis (^{MacPhee et al., 2010}); furthermore, lactobacilli have been proposed for the treatment and prevention of urinary infections, vulvovaginal candidiasis (^{Reid et al., 2009}), and even for the prevention of HIV and sexually transmitted infections (^{Bolton et al., 2008}). The mechanisms involved may include immunomodulation in the host, restoration of normal vaginal flora, and interference with pathogen colonization (^{Spurbeck and Arvidson, 2011}).

A synergy between disturbed vaginal flora or bacterial vaginosis, on the one hand, and genital HPV infection, on the other, has been documented (^{Verbruggen et al., 2006; Lippman et al., 2010; Gillet et al., 2011}).

In this preliminary study, we examine whether probiotics influence the clearance of HPV infection and associated precancerous cellular abnormalities, in casu LSIL.

Methods

Women with a diagnosis of LSIL on their latest PAP smear were eligible for participation in the study. Women with a first as well as with a repeat LSIL diagnosis were eligible for inclusion. Participants were recruited in two ways: through their gynecologist who took the PAP smear and through an informative website on HPV of the University of Antwerp. The construction and contents of the website have been described previously (^{Verhoeven et al., 2010}).

We calculated that, to obtain a 20% reduction in the chance of persistence of cytological lesions, 300 patients had to be included (150 in each group). Because we had no reference for the magnitude of a possible effect of probiotics, we decided to perform a pilot study with 50 women. Taking into account possible dropouts, 54 patients were included.

Study candidates agreed to have a new PAP smear taken in 6 months by their own physician (according to Belgian guidelines). The results of this PAP smear served as an outcome measure for the study. Liquid-based cytological samples were classified according to the Bethesda System. Participants provided three vaginal samples for HPV testing during the course of the study (t₁=<1 month after the LSIL diagnosis, t₂=3 months after diagnosis, and t₃=6 months after diagnosis). They were notified by an email when it was time to take a new swab. Samples were sent by post. The detailed procedure has been described and validated in a previous report (^{Baay et al., 2009}). Furthermore, participants filled in a 29-item questionnaire on demographic characteristics, medical antecedents, medicine use, contraceptive methods, intimate hygiene, sexual practices, number of lifetime partners, smoking and drinking habits, coitarche, health-related attitudes (physical exercise, consumption of vitamins, fruit, probiotic drinks), amount of perceived stress in daily life, and history of abnormal PAP smears and cervical cancer in family members.

Upon arrival in the laboratory, swabs were clipped into a 10 ml tube, containing 3 ml of Tris buffer (10 mmol/l Tris, pH 8.0, 1 mmol/l EDTA). The tubes were vortexed vigorously for two rounds of 10 s. The brush was removed and the tubes were centrifuged for 5 min at 3000 rpm. The supernatant was aspired and the pellet was resuspended in 0.2 ml Tris buffer. DNA was extracted using the GenElute Blood Genomic DNA Kit (Sigma-Aldrich, Bornem, Belgium), according to the manufacturers’ instructions. Briefly, cells were lysed in the presence of proteinase K and lysis solution at 55°C for 10 min, loaded onto a prepared column after the addition of 100% ethanol, washed twice, and eluted from the column with 200 µl of the elution solution provided. HPV was detected using the GP5+/6+ HPV PCR (^{De Roda Husman et al., 1995}) with a high-risk HPV probe cocktail in an enzyme immunoassay format (^{Jacobs et al., 1997}). The cut-off value for HPV positivity was calculated as the mean plus three times the SD of six negative controls in each plate. Genotyping was performed in an enzyme immunoassay with separate probes for high-risk types 16, 18, 26, 31, 33, 35, 39, 45, 51, 52, 53, 56, 58, 59, 66, 68, 73, and 82.

Blinding was not possible in this study as the individuals were informed that they were participating in a ‘probiotic’ versus ‘common care’ study, and women in the common care group were asked not to consume any probiotics during the study period. For practical reasons, we allowed participants to choose whether they wanted to be in the intervention (probiotic) or the control (common care) group. However, 39% of participants had no preference and were randomly allocated, and individuals entering the study more than 14 days after their LSIL diagnosis (24%) were automatically allocated to the control group. The intervention consisted of the daily consumption of a commercially available probiotic (Yakult) containing Lactobacillus casei Shirota (LcS), which was placed at the participants’ disposal by means of coupons by which they received the drink free of charge in a supermarket of their choice. Conforming to the standard of care, the control group received no treatment. This group was asked not to use any probiotic drinks during the 6-month study period. Both Pap smear examination and HPV detection were performed blinded to the study arm. Analyses were performed using SPSS for Windows version 18.0 (SPSS Inc., Chicago, Illinois, USA). Univariate statistics (t-test, χ², Fisher’s exact test, univariate logistic regression) were determined to compare outcomes in the intervention and control groups and to assess the impact of possible determinants. Because of the small sample size of this pilot study, we did not determine multivariate statistics.

This trial was approved by the ethical commission of Antwerp University. It was registered at http://www.clinicaltrials.gov and assigned number NCT01097356.

Results

Fifty-four participants were included in the study. Three participants did not complete the study: one emigrated after 2 months and was lost to follow-up, one became pregnant, and one decided to undergo a colposcopy instead of conservative care during the study period. Thus, we performed ‘per protocol’ analyses on 51 participants.

Table 1 shows that the basic characteristics were similar in the probiotic and the control group. Most women had a first LSIL diagnosis (first PAP smear or normal or ASCUS PAP smears in the past); five women had a repeat LSIL diagnosis for which their gynecologist decided to continue conservative treatment. For seven women, the results of their earlier PAP smears could not be retrieved. The following HPV types were detected, frequently as multiple infections: 16, 18, 31, 33, 35, 39, 51, 52, 53, 56, 58, 59, 66, and 68. Seven women in the intervention group and four women in the control group had the more virulent HPV-16 type (Table 2).

Table 1: Baseline characteristics of study participants

Table 2: Odds ratios and confidence intervals for the persistence of cytological abnormalities, with several lifestyle factors as possible determinants

Cytological results

In 39.2% of participants, cytological abnormalities had resolved: 29.6% in the common care group and 50% in the probiotic group. Figure 1 shows the PAP smear results at outcome in both groups.

Fig. 1: Study flow. ASCUS, atypical cells of undetermined significance; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion; NILM, negative for intraepithelial lesions (= cure).

The significance of ASCUS is, as the term describes, unclear. The policy in Belgium in this case is to repeat a PAP smear after 1 year. If ASCUS smears are considered as ‘no cure’, there is a tendency toward a higher chance for cure in the probiotic group [12/24 or 50.0 vs. 8/27 or 29.6%, odds ratio (OR) 2.37, P=0.137]. If, however, on the basis of the uncertainty of the diagnosis, ASCUS smears are excluded from analysis, the difference in the likelihood of resolution of cellular abnormalities between both groups is statistically significant (12/20 or 60 vs. 8/26 or 30.7%, OR 3.38, P=0.047) in favor of the probiotic group. The number of participants in this pilot is too small to perform a complete multivariate analysis; however, controlling the effect for several combinations of confounders did not substantially alter the results. More specifically, the significant effect of probiotics was still present in a model that contained the two factors in which the intervention group and the control group tended to differ: smoking and consumption of fruit. Taking into account the fact that a minority of participants was not randomized but allocated to one or another group (because they entered the study too late to be in the intervention group or because they had a strong preference for one or the other group), we checked whether there was a difference in outcome between randomized and nonrandomized participants within each group. No difference was observed.

Presence of human papillomavirus DNA

At the beginning of the study, all women were positive for HPV. After 3 months, HPV had been cleared in 16% of the study participants; in 25.0% of probiotic takers; and in 7.7% of control participants (P=0.13). In 24.0% of participants, HPV DNA was no longer detectable after 6 months; this was the case in 19.2% of the control participants and in 29.2% of probiotic takers (P=0.41).

Influence of other factors on cytological clearance

Table 2 shows the ORs and confidence intervals for persistence of cytological abnormalities, with several lifestyle factors as possible determinants. These nonsignificant tendencies are shown to compare our pilot sample with the body of known evidence on the predictors of persistence of cytological lesions (see the Discussion section, Fig. 2).

Fig. 2: Evolution of HPV positivity over time in the probiotic and the control group. HPV, human papillomavirus; t ₁, at study entry; t ₂, after 3 months; t ₃, after 6 months.

Discussion

This pilot study shows that women who took probiotics have a twice as high (60 vs 31%) chance of clearance of their HPV-infection-related cytological abnormality as a control group. Probiotics users also had higher HPV clearance rates (29 vs. 19%), but this difference was not significant.

In agreement with existing evidence, smoking tended to lower the chance of clearing the infection and the cytological lesions (^{Kjaer et al., 1996; Guarisi et al., 2009; Maucort-Boulch et al., 2010}). Similarly, genetic vulnerability (in the form of cervical cancer/abnormal cytology in the family) tends to be a risk for persistence of HPV-induced cellular abnormalities (^{Bosch et al., 2002; Wang et al., 2010}). Genetic factors are perceived as the major risk factors by patients (^{Baay et al., 2004; Ackermann et al., 2005}). Acquisition of HPV is well known to be age dependent, but once HPV has been acquired, age is not associated with clearance rates (^{Syrjänen et al., 2005}). This finding is supported by our data. Tampon use seems to be a limited risk in our study – an association with acquisition or delayed clearance has been found, but inconsistently, in the literature (^{Rousseau et al., 2000; Richardson et al., 2005}). Daily fruit consumption seems to be protective, which is consistent with several observations of a protective effect of the regular intake of fruit and fresh vegetables, which contain micronutrients that may help in preventing cellular DNA damage (^{Sedjo et al., 2002; Richardson et al., 2005}). Vitamin intake was not beneficial in our study; on the contrary, vitamin takers did markedly worse in clearing their infection. We found no evidence of a protective effect of vitamin supplementation on existing lesions in the scientific literature. Possibly, vitamin intake may be a marker of a more general lack of well-being related to a suboptimal health status. In our study, we found no tendency toward a protective effect of condom use. Although condom use is probably more important in the prevention of HPV acquisition, some studies report a protective effect of consistent condom use even after HPV infection (^{Hogewoning et al., 2003}).

Hormonal contraceptive (OAC) use tends to be protective in our sample. Although OAC has been proposed as a cofactor in cervical carcinogenesis (^{Murthy and Mathew, 2000; Castellsagué and Muñoz, 2003}), most observational studies do not find an effect on the clearance of HPV or cytological abnormalities (^{Syrjänen et al., 2006}). Some studies report a protective effect (faster clearance) with OAC, similar to our observation (^{Muñoz et al., 2002; Richardson et al., 2005}). In general, the findings in our study are in accordance with existing evidence, which supports the representativeness of our sample.

Several limitations of this pilot study need to be pointed out. There was a substantial heterogeneity among study participants (irrespective of the groups they were allocated to), especially in terms of age and (often unknown) the duration of HPV infection before entering the study. Furthermore, the fact that some of the participants chose their group instead of being randomized may have led to some form of bias, although we registered most relevant health-associated attitudes in the study questionnaire, and no significant difference could be found between the groups. The choice for taking the probiotic in some individuals might fit in with a healthy lifestyle in general, as probiotic takers tended (although not significantly) to smoke less and to eat more fruit. However, after controlling for these two factors in multivariate analysis, the significant effect of the probiotic remained. Furthermore, within each group, there was no difference in outcome for randomized versus nonrandomized participants.

Probiotics are believed, with limited evidence, to have beneficial effects on a large spectrum of diseases (infection, allergy, cancer) because of their immunomodulatory properties. The mechanisms by which probiotics act in urogenital infections are not well understood. Possible lines of defense include reconstitution of vaginal (bacterial) microflora, direct killing of pathogens, competition for host-cell receptors, and interference with gene expression of pathogens (^{Spurbeck and Arvidson, 2011}). Antiviral activity has been attributed to the secretion of specific metabolites or by influencing gene-regulatory pathways important in immune responses (^{Van Baarlen et al., 2011}). It will be a challenge to determine the in-vivo contribution of each of these proposed mechanisms toward the clinical effect. As for the impact on cancer, most evidence also comes from experimental (in-vitro and animal) studies. Several strain-specific biological effects, including normalization of gut microbiota, decrease of harmful substances produced by intestinal bacteria, and enhancement of NK-cell activity, have been suggested as possible mechanisms of anticancer activity (^{Nanno et al., 2001}). Epidemiological observations are inconclusive and intervention studies are very costly and difficult to perform, because of a long duration of follow-up and the need for a large number of participants.

Although a recent study has shown a beneficial effect (higher complete remission rate) of probiotics when added to conventional therapy on cutaneous warts (^{Cassano et al., 2011}), which are also induced by HPVs, to our knowledge, our study is the first to show an effect of probiotics on mucosal HPV and its consequences. In this study, we propose a model for examining a possible effect of a nutritional intervention on HPV infection and on LSIL abnormalities, which represent early-stage disease in the progression to cervical cancer. As LSIL often regresses spontaneously, it is not a sensitive predictor of cancer, but it has the advantage of being reasonably prevalent and to require a strict clinical follow-up, which makes research feasible. The short-term (6-month) follow-up period and the availability of well-defined biomarkers (cytological abnormalities, DNA-confirmed presence of HPV) are important advantages of this model, which should make it more feasible to prove or to negate an effect of probiotics or lifestyle interventions on this type of cancer.

Because of the small sample size and the methodological considerations mentioned above, this study cannot provide conclusive evidence on the effect of the probiotic on HPV infection and on cytological abnormalities, even if statistical significance is achieved. However, the surprisingly huge effect that was found appears to be promising and warrants further exploration. In the future, the study should be repeated with a randomized, placebo-controlled design and, if possible, inclusion should be limited to women with a first diagnosis of LSIL. Finally, it would be advisable to perform cytology in a single laboratory to account for interobserver variation in cytological diagnosis (^{Stoler and Schiffman, 2001}).

Conclusion

In this study, we used cytological and viral endpoints to study a possible effect of probiotics on cervical abnormalities. The likelihood of clearance of LSIL abnormalities was twice as high in the group of probiotics users compared with nonusers. A larger and randomized study is warranted to confirm these encouraging results.

Acknowledgements

We would like to thank Dr Jean Renard for recruiting patients and all women who voluntarily participated in our study.

This study was initiated by V.V. and partly funded by university resources. We contacted Bart Degeest, Yakult Belgium, for additional funding and for the study product. Yakult Belgium made no contribution in study design, had no access to study data, and made no contribution toward the manuscript. In an earlier large study on Yakult and respiratory tract infections in the elderly, we found no beneficial effect of the probiotic. That study has been published recently in ^{Van Puyenbroeck et al., 2012}.

Conflicts of interest

There are no conflicts of interest.