Een levonorgestrelhoudend spiraal (levonorgestrel-releasing intrauterine device; LNG-IUD)(52 mg) en endometriumablatie (EA) zijn beide behandelingen voor hevig menstrueel bloedverlies (HMB). Wat betreft effectiviteit in het bereiken van een amenorroe, het reduceren van het bloedverlies, tevredenheid en de hoeveelheid re-interventies na 1 jaar lijken ze gelijkwaardig. Beide behandelingen hebben echter voor- en nadelen. Een ablatie is meer invasief en vrouwen met een (latente) kinderwens komen er niet voor in aanmerking, omdat er tijdens zwangerschap na een ablatie ernstige complicaties kunnen optreden. Tevens werkt een ablatie niet als contraceptie; een LNG-IUD wel. Recentelijk (in 2021) is de MIRA-trial, een ZonMW doelmatigheidsstudie, gepubliceerd. Dit Nederlandse onderzoek heeft aanleiding gegeven tot een herziening van de richtlijnmodule.

1. Patient satisfaction

Patient satisfaction at three, six, twelve and 24 months

Patient satisfaction at five years

2. Reduction in blood loss

2.1. PBAC score

2.2 Amenorrhea

Amenorrhea at twelve months

Amenorrhea at 24 months

3. Quality of life

4. Surgical reinterventions

5. Recovery time

Description of studies

Rodriquez (2020) conducted a systematic review and meta-analysis to determine the effectiveness, acceptability, and safety of the LNG-IUD as compared to no treatment, placebo, or other medical or surgical therapy (for example endometrial ablation (EA)) for reducing heavy menstrual bleeding (HMB). The review systematically searched multiple databases (Cochrane Gynaecology and Fertility Specialised Register, CENTRAL, MEDLINE, Embase, PsycINFO and CINAHL) from inception to June 2019, as well as grey literature and unpublished trials in trial registers. A total of 25 RCTs were included, comprising 2,511 women of reproductive age with regular heavy periods measured either objectively or subjectively. Exclusion criteria were postmenopausal bleeding, irregular menses and intermenstrual bleeding at presentation, pathological causes of HMB, primary use of progesterone-releasing intrauterine system for any other reason than HMB, and conditions where the use of LNG-IUD is contraindicated. For the comparison relevant for our analysis, LNG-IUD versus EA, a total of eleven RCTs were included. Outcome measures included in this meta-analysis were menstrual bleeding, treatment satisfaction, quality of life, adverse events, treatment failure, requirement of surgery, and resource cost. The risk of bias for each included study was assessed by two independent reviewers using the Cochrane ‘Risk of bias’ tool.

Beelen (2021a) performed a multicenter RCT to compare the effectiveness of the LNG-IUD with EA in women with HMB. The RCT was performed between 2012 and 2016 in 26 hospitals and in a network of general practices in the Netherlands. Women were included in the study if they were 34 years or older and did not have a pregnancy wish or intracavitary pathology. A total of 645 women met the eligibility criteria, but 375 women were excluded because they declined to be randomized or participate. Finally, 270 women were included and randomly allocated to the LNG-IUD group (n=132) or EA group (n=138). Outcome measures were measured at different time points, with a maximum follow-up time of 24 months. The primary outcome was blood loss and the secondary outcomes were controlled bleeding, complications and reinterventions, amenorrhea, spotting, dysmenorrhea, presence of clots, duration of blood loss, satisfaction with treatment, quality of life, and sexual function. The mean age of women in the LNG-IUD group was 44.7 years (SD 4.6), versus 45.3 years (SD 4.9) in the EA group. Groups were largely comparable at baseline. There was a significant difference between the two groups at baseline for BMI, previous uterus surgery (myomectomy), the duration of HMB, and anticoagulants use (platelet aggregation inhibitors). There was no significant difference between the two groups for the other baseline characteristics.

Results

1. Patient satisfaction (critical)

Two studies reported the outcome measure patient satisfaction (Beelen, 2021a; Rodriquez, 2020). Beelen (2021a) measured patient satisfaction on a 5-point Likert scale at three, six, twelve, and 24 months (see Table 1), while Rodriquez (2020) described the proportion of women satisfied with the treatment at one, two, and five years of follow-up. The results of both studies are shown in Figure 1, whereby the categories in the 5-point Likert scale in Beelen (2021a) were dichotomized: ‘very satisfied’ and ‘satisfied’ were defined as ‘satisfied with the treatment’.

The risk ratio at three months was 0.54 (95% CI 0.42 to 0.69), in favour of EA. The risk ratio at 6 months was 0.69 (95% CI 0.57 to 0.82), also in favour of EA. Both RRs were considered to be clinically relevant.

The pooled risk ratio (RR) at twelve months was 0.92 (95% CI 0.85 to 1.00), in favour of EA. The pooled RR at 24 months was 0.89 (95% CI 0.82 to 0.97), also in favour of EA. Both RRs were not considered to be clinically relevant.

At five years of follow-up, de Souza (2010) (Rodriquez, 2020) reported 27 out of 27 (100%) satisfied women in the LNG-IUD group, versus 20 out of 25 (80%) satisfied women in the EA group. The RR was 1.25 (95%CI 1.01 to 1.53), in favour of LNG-IUD. This was a clinically relevant difference.

Figure 1. Proportion of women satisfied with the treatment; LNG-IUD versus EA. Note: one event is equal to a satisfied patient.

Table 1. Patient satisfaction at different follow-up times as reported in Beelen (2021a). Note: the categories in the 5-point Likert scale were dichotomized. ‘Very satisfied’ and ‘satisfied’ were defined as ‘satisfied with the treatment’, and ‘uncertain’, ‘dissatisfied’ and ‘very dissatisfied’ were defined as ‘other’.

Level of evidence

Patient satisfaction at three, twelve and 24 months

The level of evidence regarding the outcome measure patient satisfaction at three, twelve and 24 months was based on a randomized study and therefore starts at high. The level of evidence was downgraded by two levels to low because of study limitations (no blinding of patients and physicians; risk of bias, -1), and the small number of included patients (imprecision, -1).

Patient satisfaction at six months

The level of evidence regarding the outcome measure patient satisfaction at six months was based on a randomized study and therefore starts at high. The level of evidence was downgraded by two levels to low because of study limitations (no blinding of patients and physicians; risk of bias, -1), and the confidence interval crossing the lower threshold for clinical relevance (imprecision, -1).

Patient satisfaction at five years

The level of evidence regarding the outcome measure patient satisfaction at five years was based on a randomized study and therefore starts at high. The level of evidence was downgraded by three levels to very low because of study limitations (no blinding of patients and physicians; risk of bias, -1), and the very small number of included patients (imprecision, -2).

2. Reduction in blood loss (critical)

2.1 PBAC score

Eight studies included in Rodriquez (2020) reported PBAC scores a six, twelve, 24 and 36 months. As PBAC scores were reported in different ways (mean versus median) and at different time points, data were not pooled. Beelen (2021a) reported PBAC scores at three, six, twelve and 24 months. All available results will be described per time point.

At three months follow-up, Beelen (2021a) reported the number of women with a PBAC score < 75. In the LNG-IUD group, 52 out of 87 women (60%) had a PBAC score < 75, versus 90 out of 108 (83%) in the EA group. The RR was 0.72 (95% CI 0.59 to 0.87). This was considered to be a clinically relevant difference.

At six months follow-up, Barrington (2003) (Rodriquez, 2020) reported a median PBAC score of 19 (range 0 to 100) in the LNG-IUD group (n=21), versus a median of 27 (range 0 to 424) in the EA group (n=23). Both median PBAC-scores were < 75. Yet, this study must be interpreted with caution, as the menstrual PBAC scores differed between groups before they received an intervention. Beelen (2021a) reported the number of women with a PBAC score < 75 at 6 months of follow-up. In the LNG-IUD group, 70 out of 100 women (70%) had a PBAC score < 75, versus 99 out of 113 (88%) in the EA group. The RR was 0.80 (95% CI 0.69 to 0.92). This was not considered to be a clinically relevant difference.

PBAC scores at twelve months were reported in Beelen (2021a) and in five studies in the SR of Rodriquez (2020). Table 2 gives an overview of the available results per study at 12 months of follow-up. All studies except Beelen (2021a) reported a mean or median PBAC score < 75 for both groups. Beelen (2021a) reported the percentage of women in both groups with a PBAC score < 75 and found a RR of 0.98 (95% CI 0.88 to 1.08), which was not considered to be a clinically relevant difference.

Table 2. An overview of PBAC scores reported in different studies at 12 months of follow-up

At 24 months, TALIS (2006) (Rodriquez, 2020) reported a mean PBAC score of 20.6 (SD 28.8) in the LNG-IUD group (n=37), versus a mean of 75.4 (SD 91.1) in the EA group (n=31). The mean difference was -54.80 (95% CI -88.18 to -21.42).

Beelen (2021a) reported the number of women with a PBAC score < 75 at 24 months. In the LNG-IUD group, 100 out of 115 women (87%) had a PBAC score < 75, versus 124 out of 132 women (94%) in the EA group. The RR was 0.93 (95% CI 0.85 to 1.01). This was not considered to be a clinically relevant difference.

At 36 months follow-up, Kittelsen (1998) (Rodriquez, 2020) reported a median PBAC score of 7 (range 0 to 101) in the LNG-IUD group (n=19), versus a median score of 4 (range 0 to 182) in the EA group (n=22). Both median PBAC scores were < 75.

Level of evidence

PBAC score at three months

The level of evidence regarding the outcome measure PBAC score at three months was based on a randomized study and therefore starts at high. The level of evidence was downgraded by three levels from high to very low because of study limitations (no blinding of patients and physicians; risk of bias, -1), the confidence interval around the RR crossing the upper limit of clinical relevance and the small number of included patients (imprecision, -2).

PBAC score at six months

The level of evidence regarding the outcome measure PBAC score at six months was based on a randomized studies and therefore starts at high. The level of evidence was downgraded by three levels from high to very low because of study limitations (no blinding of patients and physicians and unequal baseline scores between the two groups; risk of bias, -1); conflicting results (inconsistency, -1), the wide confidence interval and ranges and the small number of included patients (imprecision, -1).

PBAC score at twelve and 24 months

The level of evidence regarding the outcome measure PBAC score at twelve and 24 months was based on a randomized studies and therefore starts at high. The level of evidence was downgraded by three levels from high to very low because of study limitations (no blinding of patients and physicians; risk of bias, -1); conflicting results (inconsistency, -1), and the small number of included patients (imprecision, -1).

PBAC score at 36 months

The level of evidence regarding the outcome measure PBAC score at 36 months was based on a randomized study and therefore starts at high. The level of evidence was downgraded by three levels from high to very low because of study limitations (no blinding of patients and physicians; risk of bias, -1); and the very small number of included patients (imprecision, -2).

2.2 Amenorrhea

Eight studies in Rodriquez (2020) reported the incidence of amenorrhea within twelve months. In the LNG-IUD group, 45 out of 210 women (21.4%) developed amenorrhea, versus 41 out of 221 (18.6%) women in the EA group. The pooled RR was 1.20 (95% CI 0.76 to 1.89; figure 2), in favour of LNG-IUD. This was not considered to be a clinically relevant difference.

Figure 2. The incidence of amenorrhea at 12 months; LNG-IUD versus EA.

Beelen (2021a) reported amenorrhea at 24 months. In the LNG-IUD group, 67 out of 115 women (58%) developed amenorrhea, versus 89 out of 132 women (67%) in the EA group. The corresponding RR was 0.86 (95% CI 0.71 to 1.05), in favour of EA. This was not considered to be a clinically relevant difference.

Level of evidence

Amenorrhea at twelve months

The level of evidence regarding the outcome measure amenorrhea at twelve months was based on randomized studies and therefore starts at high. The level of evidence was downgraded by two levels from high to low because of conflicting results (inconsistency, -1), the wide confidence interval around the pooled RR and the small number of included patients (imprecision, -1).

Amenorrhea at 24 months

The level of evidence regarding the outcome measure amenorrhea at 24 months was based on a randomized study and therefore starts at high. The level of evidence was downgraded by one level from high to moderate because of the confidence interval around the pooled RR crossing the lower threshold for clinical relevance (imprecision, -1).

3. Quality of life (important)

Quality of life (QoL) was measured in Beelen (2021a) and in five trials that were included in Rodriquez (2020). Beelen (2021a) reported the SF-36 at twelve months follow-up. Four of the studies in Rodriquez (2020) (Tam, 2006; Crosignani, 1997; Soysal, 2002; TALIS, 2006) also used the SF-36 to measure QoL at twelve months follow-up. The fifth study (de Souza, 2010) used the Psychological General Well-Being Index (PGWBI) to measure QoL at five years follow-up.

Beelen (2021a) measured the QoL at twelve months with the SF-36 and reported mean differences for all domains on the SF-36. Almost all domains showed a difference at twelve months in favour of EA: physical health (mean difference (MD) -7.50, 95%CI -13.95 to -1.05), bodily pain (MD -8.40, 95%CI -16.08 to -0.72), vitality (MD -6.10, 95%CI 11.51 to -0.69), physical role limitation (MD -8.00, 95%CI -18.34 to 2.34), general health (MD -4.20, 95%CI -10,30 to 1.90), social functioning (MD -5.60, 95%CI -11.47 to 0.27), and mental health (MD -1.50, 95%CI -5.79 to 2.79). Only one domain – emotional role limitation – showed a MD of 4.30 (95%CI -3.49 to 12.09) in favour of LNG-IUD.

Tam (2006) (Rodriquez, 2020) also reported mean differences for all domains on the SF-36. Multiple domains showed a difference at twelve months in favour of EA: general health (MD -14.40, 95%CI -22.63 to -6.17), social functioning (MD -6.70, 95%CI -12.82 to -0.58), emotional role limitation (MD -10.10, 95%CI -17.03 to -3.17), and mental health (MD -11.20, 95%CI -17.08 to -5.32), physical functioning (MD -3.10, 95%CI -11.10 to 4.90), physical role limitation (MD -2.50, 95%CI -9.31 to 4.32), and vitality (MD -5.20, 95%CI -11.46 to 1.06). Only one domain – bodily pain – showed a MD of 0.40 (95%CI -6.43 to 7.23) in favour of LNG-IUD.

Crosignani (1997) and Soysal (2002) (Rodriquez, 2020) reported median SF-36 scores per domain at twelve months. In Crosignani (1997) (Rodriquez, 2020), median scores for all eight domains were equal to or higher in the LNG-IUD group as compared to the median scores in the EA group. The same was true for Soysal (2002) (Rodriquez, 2020), except for the domain general health perception.

TALIS (2006) (Rodriquez, 2020) reported the overall SF-36 score at twelve months. The MD was 2.60 (95%CI -5.98 to 11.18), in favour of LNG-IUD.

De Souza (2010) (Rodriquez, 2020) reported an improved physical well-being (RR 1.46; 95% CI 1.11 to 1.92) and emotional well-being (RR 1.59; 95% CI 1.09 to 2.30) at five years in participants having the LNG-IUD compared to EA.

Level of evidence

The level of evidence regarding the outcome measure quality of life was based on randomized studies and therefore starts at high. The level of evidence was downgraded by three levels from high to very low because of study limitations (no blinding of patients and physicians; risk of bias, -1); conflicting results (inconsistency, -1), and the small number of included patients (imprecision, -1).

4. Surgical reinterventions (important)

Beelen (2021a) reported the number of women with surgical reinterventions (endometrial ablation and/or hysterectomy) at 24 months. Rodriquez (2020) reported treatment failure, defined as the requirement for surgery for the treatment of HMB (ablation or hysterectomy) at twelve, 24 and 36 months for four studies in total. The results at twelve and 24 months were pooled and the results from Beelen (2021a) at 24 months were added to the forest plot (Figure 3). At twelve months, the pooled RR was 1.43 (95% CI 0.63 to 3.21), in favor of EA. At 24 months, the pooled RR was 1.40 (95% CI 0.46 to 4.31), also in favor of EA. Both RR’s at twelve and 24 months were clinically relevant.

Figure 3. Surgical reinterventions at twelve and 24 months; LNG-IUD versus EA

At 36 months, Kittelsen (1998) (Rodriquez, 2020) reported that no women in the LNG-IUD (n=30) group required surgery, versus 4 women in the EA group (n=30). The risk difference was 13% (95% CI 0% to 27%).

Level of evidence

Surgical reinterventions at twelve and months

The level of evidence regarding the outcome measure surgical reinterventions at twelve and 24 months was based on randomized studies and therefore starts at high. The level of evidence was downgraded by three levels from high to very low because of conflicting results (inconsistency, -1), and the very wide confidence interval (imprecision, -2).

Surgical reinterventions at 36 months

The level of evidence regarding the outcome measure surgical reinterventions at 36 months

was based on a randomized study and therefore starts at high. The level of evidence was downgraded by three levels from high to very low because of the very small number of included patients and the very wide confidence interval (imprecision, -3).

5. Recovery time (important)

Not reported.

A systematic review of the literature was performed to answer the following question:

What are the (un)favorable effects of the levonorgestrel-releasing intrauterine device (LNG-IUD) compared to endometrial ablation as a treatment in patients with heavy menstrual bleeding without myomas/adenomyosis?

P: patients with heavy menstrual bleeding (HMB) without myomas/adenomyosis

I: levonorgestrel-releasing intrauterine device

C: endometrial ablation

O: patient satisfaction, reduction in blood loss (amenorrhea and PBAC score), quality of life (using a validated scale), reinterventions, and recovery time.

Relevant outcome measures

The guideline development group considered patient satisfaction and reduction in blood loss as critical outcome measures for decision making; and quality of life, reinterventions, and recovery time as important outcome measures for decision making.

The working group defined the outcome measures as follows:

Patient satisfaction and quality of life need to be measured with a validated scale. Recovery time is the time from start of treatment to resumption of work/daily activities. Reduction of menstrual blood loss is measured by means of the PBAC (pictorial blood loss assessment chart) score (Janssen, 1995) and amenorrhea. Amenorrhea is the absence of uterine blood loss for longer than 90 days (Fraser, 2011).

For patient satisfaction, a difference of 20% was defined as a clinically (patient) important difference. For the PBAC score the working group decided that the intervention is successful if the mean or median post-treatment PBAC score was < 75. The median scores are preferred as PBAC has a skewed distribution. A difference between groups of 25% of women with a PBAC score < 75 was considered clinically relevant. For all other outcomes, the working group used 25% for dichotomous outcome measures and 0.5 SD for continuous outcome measures as a clinically (patient) important differences (Schüneman, 2013).

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched from 2012 until January 28, 2019, and updated with published literature from 2019 to December 31, 2021. The detailed search strategy of this updated search is depicted under the tab Methods. Studies were selected based on the following criteria: systematic reviews and RCTs that compared LNG-IUD to ablation in women with HMB without myomas/adenomyosis, and that reported at least one of the following outcome measures: patient satisfaction, amenorrhea/reduction in blood loss, quality of life (using a validated scale), re-interventions, and recovery time.

The systematic literature search up to 2019 resulted in 48 hits. Initially, 22 out of 48 studies were selected  based on title and abstract screening. Two additional studies were added that were extracted from the reference lists in these articles. After reading the full text, 22 studies were excluded and two studies were included (Ghazizadeh, 2014; Lethaby, 2015)

The update search (2019-2021) resulted in 43 hits. Nine out of 43 studies were initially selected based on title and abstract screening. After reading the full text, seven studies were excluded (see the table with reasons for exclusion under the tab Methods), and two studies were included. No studies from the previous search were included (see exclusion tables).

Results

Two studies were included in the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.