Evidence for the use of Levomepromazine for symptom control in the palliative care setting: a systematic review

Publications that met the inclusion criteria were those that 1) involved individuals treated in the palliative care setting, 2) included adults, 3) evaluated pharmacological treatment of symptoms at the end of life with levomepromazine and 4) were characterized as randomized controlled trials, prospective trials, cohort studies, case series or case reports. Systematic reviews were also included but were primarily used for hand searches of references. Non-systematic or narrative reviews were excluded, but collected as a separate category as proof of existing clinical knowledge/practice. Our systematic review was limited to studies published in English or for which English abstracts were available. The period of review was from 1980 to April 2012.

The following five computerized online databases were searched in the second week of April 2012: Medline (1946 to April week 2 2012), Embase (1980 to 2012 Week 15), The Cochrane Library, PsychInfo (1806 to April week 3 2012), Ovid Nursing (1946 to April week 2 2012).

The automated search was conducted using two main components: The first component included several search terms for identification of literature relevant to palliative care, based on a master search strategy developed for that specific purpose [27], enlarged by some additional search terms. The second component contained the search terms for levomepromazine.

Search terms of the automatic search are the following:exp advance care planning/OR exp attitude to death/OR exp bereavement/OR death/OR hospices/OR life support care/OR palliative care/OR exp terminal care/OR terminally ill/OR palliat*.tw. OR hospice*.tw. OR “terminal care”.tw. OR terminally ill patient.mp. or exp terminally ill patient OR exp terminal care/OR palliat*.tw. OR hospice*.tw. OR end of life care.mp. OR EOL care.mp. OR palliative therapy.mp. or palliative therapy/OR terminally ill patient.mp. or terminally ill patient/AND levomepromazine.mp. or levomepromazine/OR methotrimeprazine.mp. OR neurocil.mp. OR nozinan.mp. OR levoprome.mp.

Table 3 shows the full electronic search strategy as performed in Embase.

After conducting the search in all databases and de-duplication, as a first step, titles and abstracts of identified studies were screened for relevance to the topic and studies considered being not relevant excluded. In a second step, full texts were sought for all studies, which appeared to meet the inclusion criteria. Conference abstracts were also included. Two independent researchers then separately reviewed all retrieved papers for relevance. Where a difference in results occurred, data was discussed and the discussion recorded. Final decisions were strictly based on adherence to the inclusion and exclusion-criteria. If agreement could not be reached, full-text analysis using a relevant quality-instrument was performed. Where there was still no agreement after thorough discussion, the study was included into the search and its relevance discussed in the publication.

Relevant studies were extracted into a qualitative synthesis table and categorised according to the following items: author, title, year of publication, journal, study design, indication for levomepromazine, study population, setting, number of study participants, number of patients under treatment, mean dose, dose range, application, measurement of effectiveness, reported adverse effects, remarks, conclusion, main results from the quality analysis process, further comments.

Studies were critically appraised and the evidence was graded based on the determinants for quality of evidence published by the Oxford Centre for Evidence-Based Medicine Levels of Evidence: Level 1: evidence from a systematic review of RCT; Level 2: evidence from a RCT; Level 3: evidence from a non-randomized controlled cohort studies Level 4: evidence from case-series or case–control or historically controlled studies Level 5: expert opinion [28]. Quality assessments were undertaken by using quality check-lists adherent to the standards gathered by the EQUATOR network [29].

A total of 33 articles involving 9 systematic reviews met the inclusion criteria for research, and reported data regarding patients treated with levomepromazine/methotrimeprazine in a palliative care setting. The search of the five databases initially provided 367 studies after de-duplication. After reviewing of titles and abstracts 270 of these papers were rejected as clearly not meeting the inclusion criteria. 19 additional records were identified through hand-searching and reference lists. Of the remaining 84 references full copies were retrieved and assessed for eligibility. Of those 48 papers met the inclusion criteria. After further examination of these studies in more detail, 33 articles remained for data extraction. 25 out of these 33 papers were found via automatic database searches and eight were found through reference tracking or hand searching. Moreover we identified 23 reviews, other than systematic or narrative, not included in this review, which could be regarded as relevant for clinical practice, and beneficially be analysed in a separate review elsewhere. Figure 1 shows a flowchart of the study selection process.

The principal characteristics of the selected articles are presented in the Additional file 1: Appendix. All relevant papers were published in peer-review journals between 1980 and 2011. Seven papers dealt with the topic of sedation, five with nausea and vomiting and one paper each with pain, delirium, several indications and side effects of levomepromazine. Regarding study design, we included six case reports, two survey studies, nine retrospective studies and seven prospective studies. Nine systematic reviews were also included.

The 12 studies and four systematic reviews on palliative sedation presented the largest group within the reviewed articles [30]–[45]. Studies concerning the use of levomepromazine/methotrimeprazine in palliative sedation varied largely in study design and sample size. In a retrospective cohort study of 29 patients sedated at home, two had their medication changed from midazolam to levomepromazine, which was effective in both patients. In one of these patients the indication for sedation was pain, in the other patient delirium was given as the rationale for use [31]. The other studies on sedation did not provide specific information on the background or indication for the use of levomepromazine in palliative sedation. The retrospective chart review of Stone et al. documented levomepromazine in 33 out of 115 reviewed patients (28,7%) in their last 5 days of life [40], Stephenson et al. found that 51–58% of sedated patients received levomepromazine according to their chart review [33]. In a survey study by Chater et al. of 100 patients, reported on by 61 selected palliative care experts, 30 (30%) had received levomepromazine for palliative sedation [39]. Sykes at al showed that levomepromazine was used in only 3 out of 114 patients (2,6%) receiving sedation in a English hospital setting, during their last week of life, but that this number increased to 30 patients (26%) in the last 24 hours [35]. Reutzel et al. asked their respondents in a retrospective survey about one case of end-stage palliative sedation during the past 12 months; in 15 out of 312 reported cases (4,8%) levomepromazine was used [34]. In a multi-center, prospective, observational study on specialized palliative care units in Japan, in a sample of 102 patients the use of levomepromazine was documented in 2 cases (1.9%) [44]. A retrospective chart review of Morita et al. demonstrated the use of levomepromazine for sedation only in 2 out of 209 patients (0,97%) [37] and in a study comparing data from 97 sedated patients in three different countries (Israel, South Africa and Spain), only one received levomepromazine [38].

One of the two case studies presented a patient suffering from motor-neurone disease in which sedation was started to allow withdrawal of mechanical ventilation [30], the other case described sedation due to intractable seizures in a patient with insular thyroid cancer and brain metastases [32].

In most of the cases presented in these studies, levomepromazine was given in combination with midazolam, only Alonso-Babarro changed to levomepromazine alone [31]. Relevant papers for the use of levomepromazine in palliative sedation with recommendations regarding doses and dose range and provided indications for sedation are shown in Table 4.

Effectiveness of sedation was measured in only some of the papers, mostly subjectively rather than with standardized tools [31, 38, 39, 41]. In the work of Alonso-Babarro, effective sedation was achieved with a level 5 or greater on the Ramsay scale and a lack of emergency calls during the process. In three studies survival was measured as main outcome criteria [37, 40, 46].

Papers concerning the use of levomepromazine for nausea and vomiting represented the second largest group: we found six studies [13, 47]–[51] and two systematic reviews [23, 52] dealing with that topic. Eisenchlas et al. reported on a sample of 70 patients with digestive cancer treated in an open-label prospective study with levomepromazine for nausea and vomiting, in which sixty patients (86%) were categorized as responders. In that study, the Pearson test revealed no association between levomepromazine dose and response to treatment, and no association between levomepromazine dose and degree of sedation [13]. In a quasi-experimental prospective study Kennett et al. showed that levomepromazine is an efficient first line antiemetic in indeterminate patho-physiological causes of nausea and vomiting, and second line for all other causes [50]. These findings were confirmed by a non-comparative prospective study conducted by Stephenson et al. in which, from a sample of 61 hospice patients with nausea, 27 (44%) received levomepromazine [48]. Moreover levomepromazine was proven to be effective as second line treatment in chemotherapy-induced nausea [51] and in carcinoid syndrome [49]. In a survey of 154 oncologists and oncology nurse prescribers levomepromazine was recommended for refractory chemotherapy-related nausea and vomiting as the second or even third-line treatment option [47]. One other paper did not focus on nausea and vomiting as a main issue, but reviewed the extent of drug use for unlicensed purposes in a palliative care unit and found that 8 out of 689 prescriptions (1,2%) were oral levomepromazine for nausea and vomiting, and 18 (2,6%) were subcutaneous levomepromazine [53]. In another systematic review on symptom management for the adult patient dying with advanced chronic kidney disease, levomepromazine was recommended as second line therapy for nausea and vomiting, if haloperidol failed. The authors do not provide doses for levomepromazine in the renal failure population [54].

Table 5 shows recommended doses and dose range of levomepromazine.

For the specific use of levomepromazine in delirious or restless patients three papers were included [56]–[58]. One retrospective chart review found that, in 39 hospital patients with delirium during their last week of life, 7 patients (18%) were treated with levomepromazine. A combined treatment of neuroleptics and benzodiazepines was used more often in that study [56]. Fainsinger at al. reported a case of agitated delirium that was treated with levomepromazine after haloperidol and lorazepam had failed. In that case the patient equally failed to respond to the doses of levomepromazine that were used (20–60 mg/24 hours), and the presence of extrapyramidal side effects contributed to the decision to change treatment again, with midazolam the final agent effectively controlling the delirium [58]. That case report is the only paper on levomepromazine included in the systematic review on the treatment of terminal restlessness performed by Kehl et al. 2004, which concluded that there is little empirical evidence suggesting that a single medication or class of medications is superior to another for terminal restlessness [57]. One final paper, which did not focus on delirium as main issue, but was reviewing the extent of drug use for unlicensed purposes in an English palliative care unit, found that 4 out of 689 prescriptions (0,6%) were subcutaneous levomepromazine [53].

The remaining three papers deal with different issues concerning the use of levomepromazine in palliative care. One study focuses on the analgesic quality of the drug. The authors report a case of a patient suffering from pain associated with lung cancer, which was sensitive to opioids and possibly related to bowel shutdown. This patient obtained adequate relief of abdominal pain with a dose of 10 mg levomepromazine i.m.. The authors used a conversion rate of 10 mg levomepromazine to 5 mg of morphine and preferred it because of a smaller effect on the gut and less respiratory depression [59]. Aside from this one case report, no further literature on the analgesic effect of levomepromazine was found. Another case report discusses the possible side effect of levomepromazine-induced lupus erythematosus in a patient with metastatic non-small cell lung cancer [60] and one conference abstract reported the use of levomepromazine in the management of terminal haemorrhage [61]. One systematic review on treatment of intractable breathlessness in patients with advanced cancer showed that there are no randomized controlled trials of phenothiazines in patients with cancer, and that the use of these agents is predicated on evidence in COPD and healthy volunteers. The authors of that review non-the-less recommend the general use of levomepromazine in patients in whom anxiety becomes overwhelming, or for palliative sedation therapy at the end of life [24].

Only 12 studies included information about effectiveness or reported information on measurement of effectiveness [13, 31, 35, 37]–[41, 48, 50, 51, 58]. In six of these studies, only specific effectiveness of the treatment with levomepromazine was provided, the summarized data is shown in Table 6.

The assessment of quality for the included studies was undertaken according to the standards gathered and regularly updated by the EQUATOR network. Risk of bias was assessed on an individual study level. The PRISMA checklist was used for the nine systematic reviews [26] and in 14 papers quality was assessed using the STROBE check-list for observational studies [62]. The six case reports included in the review were evaluated by the check-list recommended by Sorinola et al. [63], and the four papers reporting survey research were evaluated as suggested by Kelley et al. [64]. No one paper in any category covered all reporting or quality criteria as set out by their corresponding check-list. All survey studies were rated medium to high quality. The case reports averaged medium quality, and the quality of reporting for the systematic reviews and observational studies was only of low to medium quality.

Levels of evidence according to the Oxford Centre for Evidence-Based Medicine Levels of Evidence ranged from level 2b (retrospective/individual cohort study) to level 5 (expert opinion). Most papers (n = 22) were categorized as level 3 (non-randomized, non-consecutive or cohort studies). Only three studies reached level 2: one concerning palliative sedation [31] and two on nausea [13, 50]. Details of quality assessments for every study are presented in Additional file 1: Appendix. Further information about the process of quality assessment and use of the check-lists can be obtained from the authors.

As no meta-analysis was conducted and the studies included in our systematic review showed large variations of study design, sample size and quality, no assessment of risk of bias across studies was undertaken.

Because of the limitations of available studies the overall evidence for the use of levomepromazine resulting from the present review remains weak. Findings mainly based on retrospective study designs, lack of control groups, missing randomisation and small sample sizes all lead to a weak level of evidence. More homogeneous prospective studies on larger number of patients, and including measurement and reporting of outcome parameters, should be performed to provide more reliable data.

Our systematic review followed the steps considered good practice including the pre-investigation registration of our review protocol, adherence to the reporting standards and rigorous recording of decision pathways during the review process. However, some limitations apply. We did not perform any meta-analysis as the heterogeneous and low-quality data of the original studies included simply did not allow such a step, and we did not apply risk-of-bias assessment tool across studies. What is more, we limited our review to published data, deliberately excluding grey literature and non-published expert opinion, introducing a publication bias to our review.