IGSLI: Mortality and Suicide Prevention

Mortality and Suicide Prevention

Affective disorders are strongly associated with increased mortality, primarily from suicide, but also from cardiovascular events. Excess mortality in these patients is 20 to 30 times higher than in the normal population. For bipolar disorder, excess mortality is estimated to be as much as 50 to 100 times higher.

In the early 1980s Bruno Müller-Oerlinghausen former head of the Berlin Lithium Clinic (Freie Universität Berlin) became aware of the fact that the mortality in his patients was comparably low. Only 8 patients out of 200 had died over 15 years, 5 of them by documented or suspected suicide. This observation sparked intensive research efforts, including projects on the serotoninergic system and aggression. Müller-Oerlinghausen and his research group advanced the hypothesis that regular lithium prophylaxis would decrease the number of expected suicides and possibly also the number of suicide attempts. At that point in time there was only very limited data on the possible association between lithium prophylaxis and a decrease in suicidality. The first presentation of Müller-Oerlinghausen’s findings coincided with the foundation of IGSLI.

In 1992 Müller-Oerlinghausen et al. showed a significant reduction of the number of suicide attempts in lithium long-term treated patients , whereas in those patients where the medication had been interrupted by reasons whatsoever this effect was not observed. Felber et al. (1994) confirmed this finding in a well-documented patient sample from Dresden.

Replicating and broadening the initial findings from Berlin and Dresden was the aim of IGSLI’s first project. The so-called MORTA-projects were based on suitable protocols agreed upon by all IGSLI centres. Original patient data were extracted and converted into a common file related on a case-by-case basis to the corresponding national mortality statistics and analysed using the standardized mortality ratio (SMR). The first analysis showed an impressive mortality-lowering effect for lithium long-term prophylaxis in unipolar, bipolar, and schizoaffective patients (Müller-Oerlinghausen et al. 1992). Although the suicide-related SMR did not become completely normalized, i.e. was still higher than in the general population. It could be demonstrated that it was definitely lower in all diagnostic groups compared to what had to be expected in untreated patient samples.

From the very beginning of the MORTA studies it appeared very doubtful - due to ethical reasons and the low frequency of suicidal events - whether might be possible to perform a confirmatory study by including a control group comprised of affectively ill patients who were not on lithium prophylaxis. One way to corroborate the evidence from the retrospective study was to analyse data from patients who had discontinued lithium (Müller-Oerlinghausen et al. 1996), and by comparing mortality ratios from patients during initial and during later lithium treatment (Müller-Oerlinghausen et al. 1994). After discontinuing lithium treatment, patients’ mortality again rose dramatically.

At first, analysing the data based on the successive year-per-year cumulation of SMR led to the interpretation that a reduction in mortality could only be seen after several years of uninterrupted lithium treatment. However, reanalysing the data using another mathematical approach revealed this assumption to be the result of an artefact (Wolf et al. 1996). The mortality-lowering effect of lithium starts much earlier after treatment initiation.

The first publications by members of IGSLI were paralleled by a paper by Coppen et.al. (1991) who also reported a marked reduction of the standardized mortality rate down to the level of the general population in lithium long-term-treated, mostly unipolar patients. Further studies by members of IGSLI and others from various countries corroborated and differentiated these findings. Thus, Agneta Nilsson (1995) from Sweden, also a former IGSLI member, did not observe a full normalization of the SMR in a large lithium medicated cohort but also reported a marked rise of the SMR up to the expected level of untreated affective disorders when lithium had been discontinued. This observation was confirmed by another Swedish study (Kallner et al. 2000).

The latter two studies both also underline the benefit of specialized lithium clinics with optimal monitoring of the patients and their treatment. Further confirmatory studies published by non-IGSLI research groups as well as various meta-analyses of the existing data are referred to by e.g. Müller-Oerlinghausen et al. (2006).

Apart from the so far unresolved issue of the exact mechanism of the antisuicidal effect the question arises whether and to which extent it is a specific effect, in other words, is it independent of the well proven episode-preventive effect and is this effect not shared by other psychotropic agents? According to studies by members of IGSLi both question might be answered with yes. Ahrens and Mueller-Oerlinghausen (2001) based on a reanalysis of data from those IGSLI patients who had shown suicide attempts in their history before start of lithium medication could show that the reduction of the risk of suicidal events was more or less independent on whether these patients were clear-cut responders or non-responders in terms of episode-prevention. Patients may suffer from breakthrough depressive episodes in spite of ongoing lithium medication but still the risk of suicidal events is diminished. From a prospective randomized trial comparing the prophylactic efficacy of lithium vs. carbamazepine vs. amitriptylin it turns out that whereas in the group of patients treated with carbamazepin over 2.5 years 9 suicidal events, including 4 fatal ones, occurred in contrast to none in the lithium group (Thies-Flechtner et.al. 1996, Goodwin 1999). Up to now it has never been shown that long-term treatment with antidepressants or other mood stabilizers except lithium could reduce the suicide related mortality. On the contrary, a large study in the USA (Goodwin et al. 2003) reported on a 2.7 fold increased risk of suicide in valproate as compared to lithium treated patients.

It is bewildering that although the WHO Mental Health Action Plan 2013-2012 aims for a 10 %v reduction in the suicide rate and advocates the implementation of evidence-based interventions in at risk populations including people with mood disorders, it does not mention the role of lithium among pharmacological treatments….. (!) Why did the many WHO advisors overlook or suppress the clear evidence of the suicide risk lowering efficacy of lithium?

Implementing evidence in routine treatment

Affective disorders are devastating for patients and their families. This, and the high excess mortality of patients with affective disorders rates often require effective, safe, and evidence-based long-term treatment strategies. Therefore, the antisuicidal effect of lithium should be taken into account when deciding which treatment option is the most appropriate. One of the primary treatment goals in such patients clearly should be the prevention of suicide. It is, however, irritating to note that even recent therapeutic recommendations and consensus documents do not give proper consideration to this important and life-saving aspect of lithium treatment (Crossley et al. 2006).

It took more than ten years and the joint efforts of a highly motivated group of international researchers to detect the antisuicidal effect of lithium and to accumulate empirical evidence of this important effect. It is now supported by more recent data and excellent reviews of existing evidence. It will require further efforts and more years to spread the message to other countries and to implement the evidence in the routine treatment of patients with affective disorders. Suitable treatment algorithms taking the antisuicidal effect of lithium in adequate account have been proposed by members of IGSLI (Müller-Oerlinghausen et al. 2006; Berghöfer et al. 2002).

To top