Our cohort (n=55) consisted primarily of educated, white women in their late 20’s. Demographic characteristics of women with vestibulodynia did not differ significantly from the pain-free comparison group on: age (26.6 vs 27.7 yrs), ethnicity (80% vs 66.7% non-hispanic white), education (77.1% vs 50% college educated), and marital status (42.9% vs 22.2% married)
To establish reproducibility of the threshold measurement, the first measurement collected for each subject at each site was compared to the third measurement collected by the same experimenter at the same site (which was collected approximately 4 seconds later). All subjects in the study (34 vestibulodynia patients and 21 healthy controls) were included in this analysis. To assess reproducibility of the pain threshold over time, the results of the first visit were compared to those of the second visit among subjects with two visits. Sixteen women participated in two different visits, including 6 pain-free and 10 women with vestibulodynia. Similarly, when two examiners collected measurements on the same subject at the same visit, we compared the measurements of the two examiners to assess inter-examiner reproducibility. There were 16 participants for whom we collected measurements from two examiners at the same visit, including 4 pain-free and 12 women with vestibulodynia. The mean difference (and associated standard errors) and the Pearson correlation coefficients between each pair of measurements are shown in Table 1. The correlation between the first and third measurement is very high for all three types of pain threshold measurements at all sites; all such correlations except for one were 0.8 or greater. The mean differences between the first and third measurements are small (ranging from 0.018 N and 0.151 N) for the mucosal measurements. However, the third muscle measurement (for both threshold and tolerance) was consistently lower than the first measurement (the mean differences ranged between 0.55 N and 1.82 N), suggesting that sensitivity increased with repetitive stimulation. The visit 1/visit 2 correlation is very high for both mucosal and muscle measurements at all examination sites: The correlations ranged from 0.547 to 0.871 for the mucosal data and from 0.748 to 0.85 for the muscle data. The mean visit 1/visit 2 differences were modest for both mucosal and muscle measurements (ranging between 0.008 N and 0.15 N for the mucosal measurements and 0.03 N and 1.68N for the muscle measurements). The experimenter 1/experimenter 2 correlations were lower (ranging from 0.107 to 0.577), and the mean experimenter 1/experimenter 2 differences were larger (ranging from 0.071 N to 0.43 N for the mucosal measurements and 0.17 N to 3.9 N for the muscle measurements). This is not surprising given the challenges of standardizing the actual conduct of the exam (see Discussion).
Group Differences in Mucosal and Muscle Pain Sensitivity Measures
Having established the reproducibility of the pain thresholds obtained by our instruments, we evaluated their ability to discriminate between vestibulodynia patients (cases) and pain-free controls. shows Kaplan-Meier plots of the set of mucosal detection thresholds for pain-free controls and patients with vestibulodynia. There is a large difference in the distribution of the mucosal perception thresholds between patients with vestibulodynia and the pain-free comparison group. In all six of the measured mucosal sites, women with vestibulodynia showed significantly lower pressure pain detection thresholds compared to their pain-free counterparts. This was particularly robust at the lower vestibule, corresponding to sites 5, 6, and 7.
Kaplan-Meier Plots for Vulvar Mucosal Pressure Pain Detection Threshold
Similarly, compared to pain-free participants, patients with vestibulodynia showed lower pelvic muscle pressure pain threshold and tolerance measurements. shows the corresponding Kaplan-Meier plots for the muscle threshold and tolerance measurements at each of the three muscle sites.
Kaplan-Meier Plots for Pelvic Muscle Pressure Pain Threshold and Tolerance
To formally test the null hypothesis of no difference between the mucosal detection thresholds between the two groups, we fit a Cox proportional hazards model to predict a subject’s mucosal threshold based on an indicator variable for case status (see Methods). We also fit a second model that adjusted for several covariates, namely a dummy variable for set (reflecting which examiner went first and second), and an interaction term between set and case status. The effects of other covariates were not significant so they were not included in the final model. These results are summarized in .
Hazard Ratio For Termination of Pressure Application At Specified Anatomical Sites
The hazard ratio for vulvodynia case status was observed to be approximately 1.41 times the hazard ratio for controls. This difference remained basically unchanged (1.29) after adjustment for possible confounders. The hazard ratio varied from site to site. The ratio was largest at sites 5, 6, and 7 and smallest at site 12 (). This suggests that sites 5, 6, and 7 may be the most informative sites for discriminating between women with vestibulodynia and pain-free controls. Interestingly, when both examiners collected a set of measurements at the same visit, pain-free participants had a lower hazard ratio during the second set of measurements whereas vestibulodynia patients had a slightly higher hazard ratio. In other words, women without pain were less likely to terminate the procedure at a given force level during the second set than during the first set of measurements; no difference was observed in vestibulodynia patients.
A similar approach tested the null hypothesis that the vestibulodynia patients and pain-free comparison group had equal hazard ratios for the muscle measurements (both pain threshold and tolerance). Once again, we fit two Cox proportional hazards model to predict a subject’s muscle threshold based on case status as vestibulodynia (). We also fit a second set of Cox models that adjusted for covariates, but none of the coefficients of the covariates were significantly different from 0, so these results are not reported in .
We observed a significant association between vestibulodynia case status and muscle threshold measures when measurements from all three sites were combined (HR=2.9, p=0.047). Examining each site separately, there is a significant association between vestibulodynia case status and measurements at two of the three measured sites, namely sites 5 (HR=4.89, p=0.003) and 7 (HR=2.23, p<0.0001). There is a similar pattern of association with respect to the muscle tolerance measures: the measurements were associated with vestibulodynia case status across all three sites combined (HR=3.09, p=0.005), as well as at sites 5 (HR=6.31, p=0.034) and 6 (HR=4.83, p=0.049) individually.
Clinical Relevance (Validity) of the Mucosal Pain Threshold as it relates to Subjective Pain Reports
For the purpose of validating our instrument, the association between a patient’s self-reported pain during intercourse and the patient’s mucosal and muscle measurements was evaluated. Each participant was asked to rate her average pain during intercourse on a scale from 0 to 100. To assess the association between mucosal threshold and intercourse-related pain, we fit a Cox proportional hazards model to predict each woman’s mucosal threshold based on her intercourse-related pain level, as described previously. Two models were fit: one with only intercourse-related pain as a predictor, and one with intercourse-related pain adjusted for examiner and measurement set. We observed a statistically significant association between intercourse-related pain and mucosal threshold. The strength of this association varied greatly from site to site. The association was strongest at sites 6 (p=0.009), 5 (p=0.004), and 7 (p= 0.028) and much weaker and non-significant at the upper vestibular sites 2 (p=0.75), 10 (p=0.32), and 12 (p=0.38). For example, at mucosal site 6, a respondent with a mild pain report (VAS=30) has a hazard ratio of 2.46 compared to a respondent with no pain (VAS=0), and a respondent with a moderate pain report (VAS=60) has a hazard ratio of 6.05 compared to a respondent with no pain. There was no statistically significant association between intercourse-related pain and the muscle threshold or muscle tolerance measurements.
We have shown that vestibulodynia patients have lower average mucosal thresholds than pain free controls. This raises the question of whether this result can be used to more accurately diagnose patients with vestibulodynia in a clinical setting. We attempted to answer this question by using a simple procedure to develop a classifier to discriminate between vestibulodynia patients and pain-free controls. Since we previously observed that mucosal measurements at sites 5, 6, and 7 are most strongly associated with intercourse-related pain and case status, we built our classifier using only measurements from these three sites. For each subject, we averaged her first three mucosal threshold measurements at sites 5, 6, and 7. We then averaged these three averages to construct our final classifier. We only considered measurements collected in the first measurement set of the first visit for each patient. Also, we only considered measurements collected from examiner 1, since we observed considerable variation in measurements collected by different examiners, and examiner 1 collected more data than the other examiner.
The ROC curve for predicting case status based on this classifier is shown in . Note that we can obtain both high sensitivity and specificity by choosing an appropriate cutoff for discriminating between cases and controls. In particular, if we classify subjects with an average threshold force of less than 1 N as cases and subjects with an average threshold of greater than 1 N as controls, then we obtain a sensitivity of 85.7% and a specificity of 82.4%. This suggests that our methodology has the potential to help diagnose patients with provoked vestibulodynia.
Receiver Operator Curve For Predicting Case Status