Clinical effects of the levonorgestrel-releasing intrauterine device in patients with adenomyosis

Results 

Insertion of the LNG-IUD was performed without anesthesia in all cases, and no particular patient discomfort was noted. The mean age of the patients was 39.89 ± 3.91 years old, and the mean parity was 1.70 ± 0.79. There were 2 spontaneous expulsions of the LNG-IUD at 2 and 13 months after insertion, respectively. In 1 patient, the LNG-IUD was removed after the first month because of abdominal pain and irregular bleeding. These women were excluded from the postinsertion analysis. As a result, 44 patients were included in the analysis. Among these, 32 patients were followed for >36 months (Figure 1). Two patients had the original device removed, and a new LNG-IUD was reinserted at 30 months and 32 months because of increasing dysmenorrhea and vaginal bleeding.

View Large Image Download to PowerPoint

FIGURE 1. Trial profile The total number of women at admission was 47. There were 2 spontaneous expulsions of the LNG-IUD (at months 2 and 13) and 1 premature removal because of abdominal pain (at 1 month). At the end of study, there were 32 women: 10 patients participated less than 36 months but were still continuing follow-up evaluations, and 2 patients had the LNG-IUD replacement at 30 and 32 months because of increasing dysmenorrhea and vaginal bleeding. After the end of study, 15 patients had the original device removed and a new LNG-IUD was inserted. Cho. Levonorgestrel-releasing intrauterine device and adenomyosis. Am J Obstet Gynecol 2008.

The laboratory findings before and after insertion of the LNG-IUD are shown in Table 1. Serum hemoglobin, ferritin, and CA-125 levels were evaluated before and 12 months after insertion of the LNG-IUD. They all showed significant improvement at 12 months after insertion. The serum hemoglobin and serum ferritin levels increased from 11.22 ± 1.59 g/dL to 12.53 ± 2.21 g/dL (P < .001) and from 29.5 ± 23.95 ng/mL to 41.18 ± 36.11 ng/mL (P = .026), respectively; the serum CA-125 levels decreased from 44.45 ± 31.63 U/mL to 25.32 ± 16.58 U/mL (P = .009).

TABLE 1. Uterine volume, pain scores, and laboratory findings after treatment with the LNG-IUD

	Variable	Before insertion (n = 44)	6 Mo (n = 44)	12 Mo (n = 44)	24 Mo (n = 44)	36 Mo (n = 32)
	Uterine volume (mL)a	156.85±49.79b, c, d	127.17±46.85c, e	118.64±41.36b, d, e	128.84±48.70c, e	139.87±29.93c
	Pain scorea	8.55±1.02b, c, d	1.93±0.95e	1.77±0.94e	1.73±0.69b, e	3.84±1.61b, c, e
	PBAC scorea	201.43±100.17b, c, d	10.95±7.34c, e	8.45±6.72b, e	9.02±5.43e	21.55±12.26b, c, e
	Hemoglobin (g/dL)	11.22±1.59		12.53±2.21f
	Serum ferritin (ng/dL)	29.5±23.95		41.18±36.11f
	Serum CA-125 (U/mL)	44.45±31.63		25.32±16.58f

Data are expressed as mean±SD. Cho. Levonorgestrel-releasing intrauterine device and adenomyosis. Am J Obstet Gynecol 2008.

a P < .05; Friedman test. b Significant difference in comparison to the values at 6 months based on Wilcoxon signed ranks tests. c Significant difference in comparison to the values at 12 months based on Wilcoxon signed ranks tests. d Significant difference in comparison to the values at 24 months based on Wilcoxon signed ranks tests. e Significant difference in comparison to the values before insertion based on Wilcoxon signed ranks tests. f P < .05; paired t-test.

Figure 2 shows changes of the uterine volume before and after insertion of the LNG-IUD. The initial mean uterine volume was 156.85 ± 49.79 mL. After 6 months, the mean volume of the uterus significantly decreased to 127.17 ± 46.85 mL (P < .001). After 12 months, the mean volume had decreased further to 118.64 ± 41.35 mL (P < .001). Twenty-four months after insertion of the LNG-IUD, the mean volume was still significantly lower than the initial uterine volume (128.84 ± 48.70 mL; P < .001). However, the uterine volume began to increase after 12 months and was significantly higher after 36 months than after 12 months (118.64 ± 41.35 mL to 139.87 ± 29.93 mL; P = .034). The volume after 36 months was still less than the initial volume, but the difference was statistically insignificant (156.85 ± 49.79 mL to 138.05 ± 28.57 mL; P = .523).

View Large Image Download to PowerPoint

FIGURE 2. Changes in uterine volume before and after insertion of the LNG-IUD The data were analyzed with the Wilcoxon signed rank test in comparison with the preinsertion value. The most dramatic decrease in uterine volume was noted at 12 months. The volume increased steadily after 12 months; although the volume at 36 months was still less than the initial volume, the difference was statistically insignificant. A significant increase in uterine volume was noted at 36 months compared with 12 months after insertion. Cho. Levonorgestrel-releasing intrauterine device and adenomyosis. Am J Obstet Gynecol 2008.

Pain scores and PBAC scores dropped significantly after insertion of the device (Table 1). The pain scores decreased from 8.55 ± 1.02 to 1.93 ± 0.95 in 6 months (P < .001) and remained at this level after 12 and 24 months. Menstrual blood loss that was measured by PBAC scores also was reduced markedly after insertion of the device. Irregular spotting was by far the most frequent menstrual pattern in the first 6 months. However, spotting decreased progressively in subsequent months and by the end of 12 months, 10 patients (23%) became amenorrheic, 21 patients (48%) showed scanty spotting, and 13 patients (29%) had scanty but regular flows. The PBAC scores decreased significantly after the first 6 months of the insertion and represented a reduction of >90%; the reduction of PBAC scores remained similar after 24 months. As with the uterine volume, the pain scores and the PBAC scores increased after 36 months. Although they were still significantly lower compared with those of the pre-insertion period, they were significantly higher than the scores after 12 months (P = .021 and .001, respectively).

Changes in uterine artery blood flow are shown in Table 2. The RI, PI, and S/D ratio of the uterine arteries were higher 12 months after insertion. Only the PI measurements showed a significant increase in both uterine arteries (P = .002 for the right uterine artery and P = .011 for the left uterine artery). For the RI and the S/D ratios, the statistical significance was noted only in the right uterine artery. The RI and PI of both uterine arteries decreased 24 months after insertion compared with the values that were obtained at 12 months, but these differences were not statistically significant. Pearson’s correlation analysis between the uterine volume and the uterine artery blood flow are shown in Figure 3. The uterine volume displayed a significant negative correlation with the RI (Pearson’s r = −0.273, P = .001for the right; Pearson’s r = −0.246, P = .002 for the left), the PI (Pearson’s r = −0.316, P < .001 for the right; Pearson’s r = −0.190, P = .021 for the left), and the S/D ratios (Pearson’s r = −0.255, P = .002 for the right; Pearson’s r = −0.173, P = .035 for the left) of the uterine arteries.

TABLE 2. Changes in uterine artery blood flow after treatment with the LNG-IUD

	Uterine artery	Before insertion (n = 44)	12 Mo (n = 44)	24 Mo (n = 44)	36 Mo (n = 32)
	Resistance index
	Righta	0.80±0.09b	0.86±0.11c	0.82±0.07	0.81±0.05
	Left	0.82±0.08	0.86±0.15	0.84±0.09	0.83±0.08
	Pulsatility index
	Righta	1.89±0.59b	2.29±0.78c	2.03±0.56b	1.94±0.37b
	Lefta	1.95±0.56b	2.35±1.01c	2.09±0.61b	2.03±0.45
	S/D ratio
	Righta	5.34±1.98b	7.94±6.91c	6.06±2.18b	6.11±1.53b
	Left	5.59±1.88	6.06±3.11	6.26±2.88	6.40±1.97

Data are expressed as mean±SD. Cho. Levonorgestrel-releasing intrauterine device and adenomyosis. Am J Obstet Gynecol 2008.

a P < .05; Friedman test. b Significant differences in comparison to the values at 12 months based on Wilcoxon signed ranks tests. c Significant difference in comparison to the values before insertion based on Wilcoxon signed ranks tests.

View Large Image Download to PowerPoint

FIGURE 3. Correlations between uterine volume and blood flow in uterine arteries with the use of Pearson’s correlation coefficient The A, RI, the B, PI, and C, S/D ratio of both uterine arteries had negative correlations with uterine volume. Cho. Levonorgestrel-releasing intrauterine device and adenomyosis. Am J Obstet Gynecol 2008.

Comment 

The LNG-IUD is a highly effective contraceptive, and recent studies emphasize its noncontraceptive benefits, especially in treating menorrhagia and dysmenorrhea.16, 17, 18 Several studies showed a reduction in adenomyosis that was related to dysmenorrhea and menorrhagia. The reduction in adenomyosis-associated menorrhagia seems to be attributed to 2 different actions of this device: a direct effect of progestin on foci of adenomyosis and deciduation and subsequent marked hypotrophy of eutopic endometrium.6, 7, 9, 10 To our knowledge, this is the first study to evaluate the long-term effects of the LNG-IUD in patients with adenomyosis. Our results were consistent with previous work and demonstrated pain relief, reduction in the uterine volume and menstrual blood loss volume, and improvements in hematologic indices in patients with adenomyosis. It is important to note, however, that, although the uterine volume at 36 months after insertion was still lower than the preinsertion volume, the differences were not significant. The maximal reduction of the uterine volume was noted at 12 months, then it began to increase from 118.64 ± 41.35 mL to 128.84 ± 48.70 mL at 24 months. In fact, the volume that was measured at the 36-month period was significantly greater than at the 12-month period. These findings may suggest that the efficacy of the LNG-IUD on the uterine volume of patients with adenomyosis may begin to decrease 2 years after insertion.

Although little is known about the mechanism of uterine shrinkage after the uterus is enlarged by adenomyosis and then treated with an LNG-IUD, the effects on the uterus could be related to changes in uterine artery blood flow. The effect of an LNG-IUD on the uterine artery blood flow has been controversial; an increase in the impedance of blood flow in the uterine arteries by continuous intrauterine release of levonorgestrel had been suggested.9, 19, 20, 21 Our results revealed significant changes in uterine artery blood flow by LNG-IUD insertion and showed that these changes were well correlated with the changes of the uterine volume. One year after insertion, the PI from the bilateral uterine arteries and the RI of the right uterine artery increased significantly with a prominent reduction of uterine volume. However, 2 years after insertion, they started to decrease along with an enlargement in uterine volume. These findings strongly support the idea that blood flow reduction in the uterine arteries by LNG-IUD leads to uterine shrinkage.

The present study demonstrates that the LNG-IUD is effective in controlling the pain that is associated with adenomyosis. The effect of the LNG-IUD on pelvic pain has been well documented in several studies with endometriosis.22, 23 Our results were very similar, and dramatic improvements were noticed in a shorter period of time. Patients showed improvement in their symptoms starting from several months after insertion; within 6 months, most patients experience relief from dysmenorrhea. Possible explanations for the dramatic improvement of dysmenorrhea are not known, but these improvements may be related to the effects of LNG-IUD on endometrium or on the vascular supply to the pelvis with relief from pelvic congestion.

In this study, among 32 patients with follow-up periods of >36 months, 15 patients had the original device removed and a new LNG-IUD inserted. Seven patients had dysmenorrhea, and 6 patients had complaints of increased vaginal bleeding. The remaining 2 patients had no symptoms, but the size of the uterus was increasing steadily 12 months after insertion. Although the number of the patients was rather small, the statistical analysis clearly showed that the mean uterine volume became significantly larger and that the pain scores and PBAC scores began to increase significantly after 36 months of insertion in comparison with those after 12 months, along with changes in uterine artery blood flows. We do not know what exactly is responsible for these changes, but we assume that it may be related to the decline of levonorgestrel concentrations in endometrial cavity. The LNG-IUD initially releases 20 μg of levonorgestrel each day, whereas by the end of year 5 the amount drops to 14 μg every 24 hours.24 Accordingly, serum plasma concentrations of levonorgestrel in LNG-IUD users have been reported to be between 150 and 450 pg/mL in the first month after insertion and approximately 100 pg/mL by 2 years.25, 26, 27 The levonorgestrel concentrations in endometrial cavity may decrease in a similar pattern. However, because only little is known about the levonorgestrel concentration in endometrial cavity, further studies are needed to determine the levonorgestrel concentrations in endometrial cavity after long-term use of an LNG-IUD.

After 24 months, 2 patients had the LNG-IUD replaced with new one, and 10 patients continued follow-up evaluations. At the end of our study, 68% of patients had completed the 3 year-study period, which compares well with previous findings that showed a continuation rate of 56%-76% after 3 years.17, 23, 28 To date, among the 10 patients who were continuing follow-up evaluations, 4 women have completed their 3-year study period, and their findings were consistent with our result with regard to increased uterine volume, pain scores, and PBAC scores. Accordingly, we believe that the described loss of patients to follow-up evaluation probably had minimal impact on our results.

Our study had several limitations. First, our diagnostic criteria of TVS for adenomyosis had the reported sensitivity and specificity of 80%-86% and 74%-100%, respectively.11, 12, 13 Because the main reason for the use of an LNG-IUD for the management of adenomyosis was to preserve the uterus, pathologic confirmations of adenomyosis were not possible. Although TVS can be a sufficiently accurate tool for the diagnosis of adenomyosis in clinically suspicious cases, we recognize that a few false-positive or false-negative cases might have been included in this study.29 Another limitation is that this was a noncomparative study. We tried to recruit control subjects, but too many dropped out because of their symptoms and ensuing medical or surgical interventions. At the same time, it should be noted that none of the studies that involved medical treatment of adenomyosis have included control subjects. Because no other forms of medical therapy offer such long-term relief of symptoms, it would be difficult to undertake randomized, controlled, and comparable studies.30 However, further research to compare the LNG-IUD with other treatments is warranted for more definite results.

In conclusion, our results indicate that LNG-IUD is very effective in the treatment of adenomyosis with improvements of hematologic indices and symptomatic relief. Uterine shrinkage with increased blood flow impedance in the uterine arteries is evident with this treatment. However, the long-term follow-up examinations revealed a gradual increase in uterine volume, pain scores, and PBAC scores. Because these indicators start to increase 2 years after insertion with the changes of uterine artery vascularity, the efficacy of LNG-IUD may begin to decrease at this point. To maintain the efficacy of LNG-IUD for the management of adenomyosis, a new device might be needed after 3 years.