Hydergine for dementia

Abstract

Background

Currently hydergine is used almost exclusively for treating patients with either dementia, or 'age‐related' cognitive symptoms. Since the early 1980s there have been over a dozen more clinical trials, yet hydergine's efficacy remains uncertain. Although previous reviews offer generally favourable support for hydergine's efficacy, they were, however, limited by a bias with respect to the particular clinical studies chosen (e.g., the inclusion of case reports, and uncontrolled trials), and by authors' impressionistic assessments of results. Not surprisingly, there has been a lack of consensus among reviewers with regard to the efficacy of hydergine.

In 1994, a meta‐analysis was published by the present reviewers who reported that overall, hydergine was more effective than placebo. However they also observed that the statistical evidence for efficacy in 'possible or probable Alzheimer's disease' patients was so modest that one additional statistically non‐significant trial would have reduced the results to non‐significance.

Objectives

Because of uncertainty surrounding the efficacy of hydergine, the goals of this overview were to assess its overall effect in patients with possible dementia, and to investigate potential moderators of an effect.

Search methods

The trials were identified from a searches of the Specialized Register of the Cochrane Dementia and Cognitive Improvement Group, The Cochrane Library, MEDLINE, EMBASE, PsycINFO, CINAHL, LILACS, clinical trials registries and grey literature sources on 2 March 2009 using the terms hydergin*, ergoloid* and dihydroergo*. Two proprietary databases were also searched. Published reviews were inspected for further sources.

Selection criteria

Trials to be included must be randomized, double‐blind, parallel‐group, and unconfounded comparisons of hydergine with placebo for a treatment duration of greater than one week in subjects with dementia or symptoms consistent with dementia.

Data collection and analysis

Data were extracted independently by the reviewers, pooled where appropriate and possible, and the pooled odds ratios (95% CI) or the average differences (95% CI) were estimated. Where possible, intention‐to‐treat data were used.

Outcomes of interest included clinical global impressions of change and comprehensive rating scales. Potential moderating variables of a treatment effect included: inpatient/outpatient status, trial duration, age, sex, medication dose, publication year, and diagnostic grouping.

Main results

There were a total of 19 trials that met inclusion criteria and that had data sufficient for analysis.

Thirteen trials reported sufficient information to use a global rating of improvement and nine trials provided information on a comprehensive rating scale. Three trials provided both outcome measures.

It was not possible to use many of the published results in a combined analysis owing to the lack of sufficient data to perform statistical analyses.

For the 12 trials that used global ratings, there was a significant effect favouring hydergine (OR 3.78, 95% CI, 2.72 to 5.27). For the nine trials that used comprehensive ratings, there was a significant mean difference favouring hydergine (WMD 0.96, 95%CI, 0.54 to 1.37).

Hydergine was well tolerated in these trials, with 78% of randomized subjects available for data analyses.

Greater effect sizes on global ratings were associated with younger age, and possibly higher dose, although most of the subgroup analyses were statistically insignificant.

Authors' conclusions

As in an earlier systematic review, we found hydergine to show significant treatment effects when assessed by either global ratings or comprehensive rating scales (based here on a smaller set of trials than in the earlier published systematic review because trials were required to have data that could conform with MetaView, the Cochrane Collaboration statistics software). The small number of trials available for analysis, however, limited the ability of subgroup analyses to identify statistically significant moderating effects.

Unfortunately, most of the randomized, double‐blind, and placebo‐controlled trials of hydergine were conducted and published before the advent of consensus‐based diagnostic standards of dementia in 1984; therefore diagnostic criteria were less specific. As a result, uncertainty remains regarding hydergine's efficacy in dementia.

Plain language summary

Uncertainty about the efficacy of hydergine in dementia

Hydergine has been used to treat patients with either dementia, or 'age‐related' cognitive symptoms. Hydergine may offer benefit to some patients, possibly at doses of 4.5 to 9.0 mg per day, with possibly greater benefit to younger subjects and inpatients. The statistical evidence for efficacy in 'possible or probable Alzheimer's disease' patients was so modest however that one additional statistically non‐significant trial would have reduced the results to non significance; evidence for efficacy in vascular dementia rested on relatively stronger effects for hydergine on clinical ratings; and effective doses may be higher than 3 mg/d (i.e., than that currently approved in the USA). Despite its availability for 40 years, the circumstances of hydergine's efficacy in dementia syndromes have not been adequately researched, and have yet to be precisely defined.

Background

Hydergine, the brand name for a specific combination of four dihydro derivatives of ergotoxine, also referred to as ergoloid mesylates, was introduced to clinical medicine in 1949. It has been used in the past as a treatment for peripheral vascular disease, hypertension, angina pectoris, and tinnitus. Currently, it is used almost exclusively for treating patients with either dementia, or 'age‐related' cognitive symptoms. In the United States, the Food and Drug Administration has approved hydergine for 'idiopathic decline in mental capacity'. In some countries hydergine is marketed as a non‐disease‐specific cognitive enhancer.

Of note, the drug has also been advocated as a so‐called 'smart drug', for use by cognitively intact young and older adults to 'increase mental abilities'.

As early as 1981, Jarvik 1981 identified limitations in the hydergine medical literature, including the lack of diagnostic rigor, and the use of crude outcome instruments, finding that there was "no convincing consistent improvement in memory, learning, or other cognitive functions." Yet, she also observed that statistically significant improvement in confusion was reported in approximately 50% of studies. Jarvik 1981 concluded that hydergine "may help some patients with some of their activities of daily living, some of their symptoms, and their self‐care...But, overall, the improvements...are very small..."

Since that time there have been over a dozen more clinical trials, yet hydergine's efficacy remains uncertain. Although previous reviews offer generally favorable support for hydergine's efficacy, they were, however, limited by a bias with respect to the particular clinical studies chosen (eg, the inclusion of case reports, and uncontrolled trials), and by authors' impressionistic assessments of results. Not surprisingly, there has been a lack of consensus among reviewers with regard to the efficacy of hydergine.

In 1994, a meta‐analysis was published by the present reviewers (Schneider 1994) who reported that overall, hydergine was more effective than placebo as assessed by comprehensive ratings, clinical global ratings, and combined cognitive measures. Inpatient status, high daily doses, and vascular dementia were generally associated with larger effect sizes. The effect in patients who likely would have fulfilled criteria for 'possible or probable Alzheimer's disease' (ie, DSM III or NINCDS‐ADRA criteria) was significant only for combined neuropsychological measures in the five trials identified.

The present reviewers observed that the statistical evidence for efficacy in 'possible or probable Alzheimer's disease' patients was so modest that one additional statistically non‐significant trial would have reduced the results to non significance; evidence for efficacy in vascular dementia rested on relatively stronger effects for hydergine on clinical ratings; and effective doses may be higher than 3 mg/d (ie, than that currently approved in the USA).

Objectives

The aim of this review is to assess the overall effect of hydergine in patients with dementia and to investigate variables that might moderate the effect.

Additional reasons to undertake this review are to test for significant consensus among trials that may have contradictory findings; to gather potential information about efficacy which can be revealed only by assessing systematic variations in study design, data characteristics, and methodology; and to assess the development of a particular research domain.

Methods

Criteria for considering studies for this review

Types of studies

Studies were selected for this review if they fulfilled the following criteria:

(1) they comprised a clinical trial 
 (2) the trial was double‐blind, parallel‐group, placebo‐controlled, with randomized and unconfounded treatment assignment to placebo or hydergine 
 (3) sample selection criteria were specified 
 (4) subjects included elderly patients with clinical descriptions or diagnostic syndromes consistent with dementia (this criterion is intended to accommodate varying definitions of dementia over the last 30 years and to ensure adequate description of study samples) 
 (5) absence of other psychiatric disorders 
 (6) outcome instruments were specified 
 (7) sufficient statistical information was provided to calculate an effect size between drug and placebo treatment groups using Cochrane Review Manager software (version 3.0)

Types of participants

Elderly patients with possible dementia, as evidenced by clinical descriptions or diagnostic syndromes consistent with dementia (e.g., cerebral insufficiency, vascular dementia, cerebral deterioration, senile dementia, 'possible or probable AD' (NINCDS‐ADRA or DSM III criteria), presence of confusion or memory impairment, absence of other psychiatric disorders).

Two reviewers independently assigned clinical trials into five categories based on the selection of subjects as follows: 
 (1) those including patients judged to have dementia consistent with possible Alzheimer's disease (AD) 
 (2) those including patients judged to have dementia consistent with possible vascular dementia 
 (3) trials which included patients whose syndromes were consistent with a primary dementia but of undetermined type 
 (4) trials with patients described as having "cerebral insufficiency" 
 (5) and trials with patients described as having "cerebral or senile deterioration."

Categories (4) and (5) were operational definitions in the literature from the late 1960s to the early 1980s.

Types of interventions

1) Any oral dose of hydergine (including sublingual, oral tablet, or liquid capsule) 
 2) Placebo with or without multivitamin supplement

Types of outcome measures

1. Comprehensive rating scores. These typically consisted of multi‐item rating scales (eg, Sandoz Clinical Assessment Geriatric; Shader 1974). 
 2. Clinical global impressions. These typically were seven‐point clinician's ratings of change (eg, Clinical Global Impression of Change; Guy 1976). 
 3. Neuropsychological tests.

Search methods for identification of studies

See Cochrane Dementia and Cognitive Improvement Group methods used in reviews.

The Specialized Register of the Cochrane Dementia and Cognitive Improvement Group (CDCIG) was searched on 2 March 2009. This register contains records from the following major healthcare databases: The Cochrane Library, MEDLINE, EMBASE, PsycINFO, CINAHL and LILACS, and many ongoing trial databases and other grey literature sources. The following search terms were used: hydergin*, ergoloid* and dihydroergo*.

The Cochrane Library, MEDLINE, EMBASE, PsycINFO, CINAHL, many ongoing trial databases and other grey literature sources were searched separately on 2 March 2009 for the latest records. The search terms used to identify relevant controlled trials on dementia, Alzheimer's disease and mild cognitive impairment for the Group's Specialized Register can be found in the Group's module on The Cochrane Library. These search terms were combined with the following search terms and adapted for each database, where appropriate: hydergin*, ergoloid* and dihydroergo*. See Appendix 1 for the sources searched and number of hits retrieved.

Further trials were identified by Sandoz Pharmaceutical Corporation in 1991, who provided lists of published and unpublished reports under their sponsorship. These included reports supporting an original submission to a National Academy of Sciences‐‐Medical Research Panel in 1974, and a subsequent submission to the Drug Safety and Efficacy Study Implementation.

Published reviews were inspected for additional sources.

Data collection and analysis

• Selection of studies

A single reviewer (JO) discarded irrelevant citations, based on the title of the publication and its abstract. Any suggestion that an article could possibly be relevant, caused it to be retrieved for further assessment.

Two reviewers (AL and SL) independently selected the trials for inclusion in the review from the culled citation list. Disagreements were resolved by discussion by a third reviewer (JO).

• Quality Assessment

The same two reviewers (AL and SL) assessed the methodological quality of each trial using the Cochrane Collaboration guidelines (Mulrow 1996).

Category A (adequate) is where the report describes allocation of treatment by: (i) some form of centralized randomized scheme, such as having to provide details of an enrolled participant to an office by phone to receive the treatment group allocation; (ii) some form of randomization scheme controlled by a pharmacy; (iii) numbered or coded containers, such as in a pharmaceutical trial in which capsules from identical‐looking numbered bottles are administrated sequentially to enrolled participants; (iv) an on‐site or coded computer system, given that the allocations were in a locked, unreadable file that could be accessed only after inputting the characteristics of an enrolled participant; or (v) if assignment envelopes were used, the report should at least specify that they were sequentially numbered, sealed, opaque envelopes; (vi) other combinations of described elements of the process that provides assurance of adequate concealment.

Category B (intermediate) is where the report describes allocation of treatment by: (i) use of a "list" of "table" to allocate assignments; (ii) use of "envelopes" or "sealed envelopes"; (iii) stating the study as "randomised" without further detail.

Category C (inadequate) is where the report describes allocation of treatment by: (i) alternation; (ii) reference to case record numbers, dates of birth, day of week, or any other such approach; (iii) any allocation procedure that is entirely transparent before assignment, such as an open list of random numbers or assignments.

Empirical research has shown that lack of adequate allocation concealment is associated with bias. Trials with unclear concealment measures have been shown to yield more pronounced estimates of treatment effects than trials that have taken adequate measures to conceal allocation schedules, but less pronounced than inadequately concealed trials (Chalmers 1983; Schulz 1995). Thus trials are included if they conform to categories A or B, and those falling into category C were excluded.

Other aspects of trial quality are not assessed by a scoring system although details were noted of blinding, whether intention‐to‐treat analyses were extractable from the published data, and the number of patients lost to follow‐up.

• Data extraction

Data were independently extracted by the same two reviewers and cross‐checked by a third reviewer (AN). Any discrepancies were discussed.

For comprehensive rating scales, data were required to provide either: (a) means and standard deviations for pre‐ and post‐tests, (b) means and standard deviations for change scores, (c) individual data for each patient in the trial. For global ratings, data were required to provide either: (a) the percentage of responders for both drug and placebo groups, (b) frequencies of responders for both drug and placebo groups, (c) individual data for each patient in the trial.

In studies where a crossover design was used, only data from the first treatment period were eligible for inclusion.

• Data analysis

For continuous or ordinal variables, such as comprehensive rating scales, the main outcome of interest was the change in score from baseline (i.e. pre‐randomisation or at randomisation) to the final assessment. If ordinal scale data appear to be approximately normally distributed or if the analysis that the investigators performed suggested parametric tests were appropriate, then the outcome measures were treated as continuous data.

For binary outcomes such as global ratings, the endpoint itself is of interest and the Peto method of the 'typical odds ratio' used.

A test for heterogeneity of the treatment effect between the trials was made using a standard chi‐squared statistic. If a test of heterogeneity was negative then a weighted estimate of the typical treatment effect across trials, the 'typical odds ratio' (i.e. the odds of an unfavourable outcome amongst treatment‐allocated patients to the corresponding odds amongst controls) was calculated using Peto's log‐rank test adapted for ordinal data (EBCTCG, 1990). If, however, there was evidence of heterogeneity of the treatment effect between trials then either only homogeneous results were pooled, or a random effects model used (in which case the confidence intervals would be broader than a fixed effects model).

Additional hypotheses tested were that hydergine had no differential effect, when compared with placebo, for certain subgroups of patients:

1. Age (stratified into 65‐75 and 75‐85 years). 
 2. Sex (stratified into < 51% male, >50% male). 
 3. Diagnosis (Alzheimer's Disease, Vascular Dementia, Primary Dementia, Cerebral Deterioration, Cerebral Insufficiency). 
 4. Dose (1.5‐3.0mg/d and 4.0‐9.0mg/d). 
 5. Duration of treatment (<10 weeks, 12‐16 weeks, >23 weeks). 
 6. Inpatient or outpatient status. 
 7. Publication year (stratified into 1955‐1973 and 1974‐1990).

Results

Description of studies

Nineteen trials were identified which met the inclusion criteria for the review, and provided sufficient outcome data for analysis using the Cochrane Review Manager software.

• Methods

All 19 trials were parallel group designs. Treatment lasted from 9 weeks to 15 months.

• Participants

The number of participants randomised in the trials ranged from 19 to 458, with completers ranging from 19 to 274.

• Interventions

Eighteen trials were designed to compare hydergine with placebo. One trial compared hydergine to placebo with a multivitamin and 600mg of potassium (t.i.d.). The dose of hydergine and method of dose determination varied widely. In some trials, different doses of hydergine were used within the same study. Eighteen trials tested a single dose of hydergine. One trial tested 1.5mg per day, 7 trials tested 3mg per day, 6 trials tested 4.5mg per day, 2 trials tested 6mg per day, 1 trial tested 7.5mg per day, and 2 trials had different fixed doses within the same study. One trial (Hollister 1955) started at 4mg per day for 30 days, reducing the dose to 3mg per day for 15 days and 2mg per day for the final 15 days. Another trial (van Loveren 1984) started at 4.5mg per day for 12 weeks and increased the dose to 7.5mg per day for the remaining 12 weeks.

• Outcome measures

The outcome measures fall within the general groupings of (1) comprehensive rating scales, (2) clinical global impressions and (3) neuropsychological tests scores. Neuropsychological tests were not used as a separate outcome because there were insufficient trials reporting data for analysis (N=7). Within these groups there is a wide variety of scales. A few of the better known ones were used in the majority of trials, but there are many scales which are less well known and which display varying degrees of comparability whilst purporting to measure the same outcome. The number of subjects who failed to compete the trial, and of subjects who were omitted from analyses were reported.

The following outcomes were recorded:

1) Comprehensive Rating Scales 
 SCAG ‐ Sandoz Clinical Assessment Geriatric (n=8) 
 MACC Behavioral Adjustment Scale (n=1)

2) Scores of Global Function 
 Unnamed Global Rating Scale (n=9) 
 SCAG global rating (n=1) 
 Proportion based on composite of all measures administered (n=1)

3) Neuropsychological Tests 
 Benton Visual Retention Test (n=1) 
 Buschke Selective Reminding Test (n=3) 
 Digit Symbol Substitution Test (n=1) 
 Divided Attention Task (n=1) 
 Labyrinth Test (n=1) 
 Mini‐Mental State Examination (n=2) 
 Orientation (n=1) 
 Nowlis Test (n=1) 
 Raven's Progressive Matrices (n=1) 
 Toulouse‐Pieron Test (n=1) 
 Trail Making Test (n=3) 
 Wechsler Adult Intelligence Scale (n=1) 
 Wechsler Block Design (n=1) 
 Wechsler Digit Span (n=2) 
 Wechsler Digit Symbol (n=4) 
 Wechsler Memory Scale Logical Memory (n=1) 
 Wechsler Memory Scale Visual Reproduction (n=1) 
 Word Recall (n=1)

Risk of bias in included studies

Most of the included trials have significant methodological limitations by current standards, primarily due to sample selection. Eleven of the included trials were published prior to 1980, with one paper published as early as 1955. Only one trial used NINCDS‐ADRA criteria for AD. All trials were parallel group designs, and most subjects (77.9%) were able to complete the protocol.

Effects of interventions

There were a total of 19 trials that met inclusion criteria and that had data sufficient for analysis. Trials were published between 1955 and 1990 (modal year =1973).

There were 12 trials that provided data on global improvement. The analyses gave statistically significant results in favour of treatment (OR 3.78; 95%CI 2.52 ‐ 5.27), though only 7 of the trials showed statistically significant effects. In addition, the trials were significantly heterogeneous.

Nine trials that provided data regarding change on comprehensive rating scales were significantly heterogeneous. The analyses gave statistically significant results in favour of treatment (WMD 0.96; 95%CI 0.54 ‐ 1.37), though only five trials showed statistically significant effects.

SUBGROUP ANALYSES

• Age

Trials were divided into two groups based on younger (65‐75 years) and older (76‐85 years) subjects. Using global ratings, the four trials with younger subjects showed a greater effect (OR 7.72; 95%CI 3.41 ‐ 17.52) than the seven trials with older subjects (OR 3.09; 95%CI 2.11 ‐ 4.52), though the difference was not statistically significant. Using comprehensive rating scales, a similar finding occurred with the three trials with younger subjects showing a greater effect (WMD 1.48; 95%CI ‐0.05 ‐ 3.01), than the three trials with older subjects (WMD 0.83; 95%CI 0.40 ‐ 1.27), though the difference was non‐significant, and the trials with younger subjects, the treatment effect was statistically significant.

• Sex

Trials were divided into two groups based on the presence of male subjects (0‐50%, 51‐100%). For global ratings, the seven trials with fewer male subjects showed a similar effect (OR 3.65; 95%CI 2.37 ‐ 5.63) to the four trials with more male subjects (OR 3.61 95%CI 2.02 ‐ 6.46). For comprehensive rating scales, the six trials with fewer male subjects showed a smaller effect (WMD 0.90; 95%CI 0.48 ‐ 1.32) than the two trials with more male subjects (WMD 2.66; 95%CI ‐0.41 ‐ 5.74) though the difference was not statistically significant, and the latter mean effect was not statistically significant.

• Diagnosis

The trials were divided into five diagnostic groups (AD, Vascular Dementia, Primary Dementia, Cerebral Deterioration, Cerebral Insufficiency). Using global ratings, the three trials with primary dementia diagnoses (OR 7.46; 95%CI 3.35 ‐ 16.54) showed the greatest effect size, followed by the four trials with cerebral insufficiency diagnoses (OR 4.65; 95%CI 2.74 ‐ 7.89), the three trials with cerebral deterioration diagnoses (OR 3.07; 95%CI 1.67 ‐ 5.63), and the two trials with vascular dementia (OR 1.43; 95%CI 0.58 ‐ 3.49). None of the differences among diagnostic groupings was statistically significant, and the effect for the vascular dementia trial was nonsignificant. Using comprehensive ratings, the two trials with AD diagnoses did not show a statistically significant treatment effect (WMD 3.04; 95%CI ‐0.39 ‐ 6.47), though the effect was greater than in the four trials with primary dementia diagnoses (WMD 2.16; 95%CI ‐0.58 ‐ 3.75) and the three trials with cerebral deterioration diagnoses (WMD 0.83; 95%CI 0.40 ‐ 1.27).

• Dose Response

Studies were divided into two groups based on low (1.5‐3.0mg/d) and high (4.0‐9.0mg/d) daily dose. Using global ratings, the 8 trials with higher doses had a greater effect (OR 4.39; 95%CI 2.99 ‐ 6.45) than the 4 trials with lower doses (OR 2.48; 95%CI 1.29 ‐ 4.75), though the difference was not statistically significant. Using comprehensive rating scales, the 2 trials with lower doses (WMD 2.52; 95%CI ‐0.37 ‐ 5.42) showed a greater effect than the 7 trials with higher doses (WMD 0.93; 95%CI 0.51 ‐ 1.34), though the difference was not statistically significant. In addition, the trials with lower doses did not show an overall statistically significant effect.

• Duration

Trials were divided into three groups: short (< 10 weeks), medium (12‐16 weeks), and long (> 23 weeks). There were no significant differences between groupings, with the 6 trials of medium duration showing similar efficacy (OR 4.07; 95%CI 2.40 ‐ 6.88) to the four trials of long duration (OR 3.86; 95%CI 2.44 ‐ 6.09). The two trials of short duration did not show a statistically significant effect (OR 2.28; 95%CI 0.69 ‐ 7.55). Using comprehensive ratings, there was only one short duration trial compared with five medium duration (WMD 2.26; 95%CI 0.93 ‐ 3.58) and three long duration trials (WMD 0.79; 95%CI 0.35 ‐ 1.22). The trials of medium duration showed a statistically significant difference from trials of long duration.

• Inpatient Status

Trials were divided into an inpatient and outpatient group. Using global ratings, all ranges showed significant effects. There was only one outpatient trial compared to nine inpatient trials (OR 3.46; 95%CI 2.30 ‐ 5.19). Using comprehensive ratings, the three inpatient trials showed a greater effect (WMD 1.46; 95%CI 0.34 ‐ 2.57) compared with the four outpatient trials (WMD 0.85; 95%CI 0.40 ‐ 1.29); though the difference was not statistically significant.

• Publication Year

Trials were divided into two groups based on year of publication: early (1955‐1973) and late (1974‐1990). Using global ratings, all ranges showed significant effects. The early trials showed a similar effect (OR 3.76; 95%CI 2.33 ‐ 6.06) to the late trials (OR 3.81; 95%CI 2.41 ‐ 6.03). Using comprehensive ratings, there was only one early trial compared with eight late trials (WMD 0.91; 95%CI 0.49 ‐ 1.33); thus, a subgroup analysis was made.

Adverse events were not reported in a systematic manner in most of the trials so making formal comparison difficult. Seventy‐seven percent of the subjects completed the trials and had data available for analysis.

Discussion

In the two rating areas assessed, the results of this overview seem to demonstrate a heterogeneous, statistical treatment effect for hydergine, consistent with our earlier report, though about half of the trials individually showed nonsignificant effects. In particular, our results suggest that hydergine may be more likely to benefit younger subjects rather than older and inpatients rather than outpatients, though these findings were not statistically significant. In addition, higher doses may offer a greater benefit.

To maintain an adequate level of methodological rigor, only trials that were double‐blind, placebo‐controlled, and parallel‐group in design were included in this overview. In addition trials were required to report sufficient data to use the Cochrane Collaboration software. Despite these selection criteria, individual trials still may have contained methodological problems.

For example, historical factors may have biased findings in favour of hydergine: Most of the trials were published prior to 1980 and earlier trials tended to report their outcomes more selectively. The use of analyses ‐ wherein only subjects who complete a trial are analysed for efficacy ‐ in the vast majority of published trials, may have biased the outcomes toward greater effects for hydergine because dropouts may have been less likely to benefit from treatment. Historical factors in the rendering of diagnoses could have biased effect sizes because the majority of trials were reported before stringent diagnostic standards were in widespread use. 
 
 Apart from methodological biases that could positively skew effect sizes, these analyses, by design, provide no direct information about onset or duration of action within each trial; nor do they provide information on direction of change over time for hydergine or placebo. Thus, it remains unclear whether the hydergine group actually improved, or merely deteriorated more slowly than the placebo group.

The small sample sizes of most trials (Mean, 67 subjects; range, 19‐274) contributed to the frequency of statistically nonsignificant trials. The statistical power of the average‐sized hydergine trial in our earlier published review suggested that 144 subjects would be needed to ensure a type II error rate of 20 percent. Even a trial of this size may have been insensitive to cognitive change in dementia, as evidenced by current trials of drugs considered more potent than hydergine.

A general 'publication bias', whereby trials with significant results tend to be reported and those with nonsignificant results do not, may have increased the reported effect sizes. This becomes a critical issue in any meta analysis where the effect sizes are modest in magnitude. An effort was made, however, to identify all trials, including unpublished ones. Three unpublished trials were included in these analyses. These trials largely resulted in nonsignificant outcomes, tending to give a more conservative estimate of hydergine's effect.

Hydergine in possible Alzheimer's and vascular dementia

None of the Alzheimer's trials included a global rating. The two trials with comprehensive rating scales showed a nonsignificant effect. On the other hand, the two vascular dementia trials did not have a comprehensive rating that could be analysed and when global ratings were used, the effect was again equivocal. These results are difficult to compare with our earlier review because of the smaller number of trials.

A total of only 94 subjects were included in these two trials, much less than the number of subjects in typical, current multicenter trials for an Alzheimer's drug, and about a third the number from our earlier review. Clearly, the effect of hydergine in patients with possible Alzheimer's dementia has not been adequately researched.

An adequately designed clinical trial would need to include over 200 moderately impaired subjects, require a dose ranging between 4.5 and 9.0 mg/day (ie, greater than the approved dosage in the United States), track outcome over a duration of at least 24 weeks, and a relevant neuropsychological test battery.

Conclusions

Little evidence for the efficacy of hydergine in placebo‐controlled, parallel‐group trials rests on observations that hydergine shows statistical effects on global and comprehensive measures, and the suggestion that effects are stronger with higher doses. The apparent dose‐response relationship (ie, doses greater than 4.5 mg/day were more effective than lower doses) suggests that an adequate hydergine dose may be higher than the dose approved by the Food and Drug Administration.

There was little evidence that hydergine's effect varied by diagnosis, owing, probably, to the small number of trials.

Because hydergine's effects were significant at higher doses, it may be a useful comparator medication in clinical trials with new, investigational drugs with putative cognitive activity (as has been done in one study previously). Overall, patients with a primary dementia show effects with hydergine. Yet, despite its availability for 40 years, the circumstances of hydergine's efficacy in dementia syndromes have not been adequately researched, and have yet to be precisely defined.

Authors' conclusions

Implications for practice.

Hydergine showed a significant treatment effect compared with placebo. The results of this review suggest that hydergine may offer benefit to some patients, possibly at doses of 4.5 to 9.0 mg per day, with possibly greater benefit to younger subjects and inpatients.

Implications for research.

Despite the large number of trials performed using hydergine, future trials are needed in clinically diagnosed subjects with Alzheimer's disease or vascular dementia. In particular, trials using higher doses of the medication (e.g., 9 mg/d) are warranted. Finally, studies that assess long‐term use of hydergine and its ability to affect the course of disease would be appropriate.

What's new

Date	Event	Description
27 March 2009	New search has been performed	Update searches were run on 2 March 2009, and no new studies were found for inclusion in the review; a number of studies have been excluded

History

Protocol first published: Issue 2, 1997
 Review first published: Issue 2, 1998

Date	Event	Description
20 June 2008	Amended	Converted to new review format.
3 November 2003	Amended	November 2003: Jason Olin has withdrawn as author of this review due to a new conflict of interest; we are presently recruiting a reviewer to take on the update of the review.
28 March 2000	New citation required and conclusions have changed	Substantive amendment

Appendices

Appendix 1. Sources searches and hits retrieved

Source	Date Range Searched	Hits Retrieved
Medline (Pubmed)	Jan 2000‐ 2 March 09	10
Embase (Ovid SP)	Jan 2000‐ 3 March 09	13
PsycInfo (Ovid SP)	Jan 2000‐ 3 March 09	0
Cinahl (Ovid SP)	Jan 2000‐ 3 March 09	1
Lilacs (bireme)	Jan 2000‐ 2 March 09	0
CDCIG SR*	Searched 2 March 09 from 2000 onward	9
CENTRAL (The Cochrane Library)	Issue 1 2009	13
ISTP Conference Proceedings http://portal.isiknowledge.com/portal.cgi	Jan 2000‐ 2 March 09	5
Australian Digital Theses Program http://adt.caul.edu.au/	Searched 3 March 2009	0
Canadian Theses and Dissertations http://www.collectionscanada.ca/thesescanada/index‐e.html	Searched 3 March 2009	0
WHO trials register	Searched 3 March 2009	0
Current Controlled trials: Meta Register of Controlled trials (mRCT) http://www.controlled‐trials.com/	Searched 3 March 2009	0
ISRCTN Register	Searched 3 March 2009	0
Nederlands Trial Register http://www.trialregister.nl/trialreg/index.asp	Searched 3 March 2009	0
ClinicalTrials.gov http://www.ClinicalTrials.gov	Included in WHO portal	//
IPFMA Clinical Trials Register www.ifpma.org/clinicaltrials.html	Searched 3 March 2009	0
UMIN Japan Trial Register http://www.umin.ac.jp/ctr/	Searched 3 March 2009	0
OPENsigle	Searched 3 March 2009	0

Data and analyses

Comparison 1. Hydergine versus Placebo: Global Rating (% improved).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Patient status	9	393	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.46 [2.30, 5.19]
1.1 Inpatient	9	393	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.46 [2.30, 5.19]
1.2 Outpatient	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Diagnosis	11	617	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.64 [2.57, 5.14]
2.1 Alzheimer's Disease	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Vascular Dementia	2	78	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.43 [0.58, 3.49]
2.3 Primary Dementia	2	61	Peto Odds Ratio (Peto, Fixed, 95% CI)	9.66 [3.11, 30.02]
2.4 Cerebral Deterioration	3	175	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.07 [1.67, 5.63]
2.5 Cerebral Insufficiency	4	303	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.65 [2.74, 7.89]
3 Mean Age	11	617	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.64 [2.57, 5.14]
3.1 65‐75 Years	4	102	Peto Odds Ratio (Peto, Fixed, 95% CI)	7.72 [3.41, 17.52]
3.2 75‐85 Years	7	515	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.09 [2.11, 4.52]
4 Gender	11	617	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.64 [2.57, 5.14]
4.1 0‐50% Male	7	339	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.65 [2.37, 5.63]
4.2 51‐100% Male	4	278	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.61 [2.02, 6.46]
5 Treatment length (weeks)	11	617	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.64 [2.57, 5.14]
5.1 Short (< 10 weeks)	2	54	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.28 [0.69, 7.55]
5.2 Medium (12‐16 Weeks)	6	232	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.07 [2.40, 6.88]
5.3 Long (> 23 weeks)	3	331	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.57 [2.17, 5.88]
6 Mean Daily Dose	11	617	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.64 [2.57, 5.14]
6.1 Low (1.5‐3.0mg/d)	3	111	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.64 [0.74, 3.62]
6.2 High (4.0‐9.0mg/d)	8	506	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.39 [2.99, 6.45]
7 Publication year	11	617	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.64 [2.57, 5.14]
7.1 Early (1955‐1973)	8	288	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.76 [2.33, 6.06]
7.2 Middle (1974‐1990)	3	329	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.51 [2.13, 5.80]

1.1. Analysis.

Comparison 1 Hydergine versus Placebo: Global Rating (% improved), Outcome 1 Patient status.

1.2. Analysis.

Comparison 1 Hydergine versus Placebo: Global Rating (% improved), Outcome 2 Diagnosis.

1.3. Analysis.

Comparison 1 Hydergine versus Placebo: Global Rating (% improved), Outcome 3 Mean Age.

1.4. Analysis.

Comparison 1 Hydergine versus Placebo: Global Rating (% improved), Outcome 4 Gender.

1.5. Analysis.

Comparison 1 Hydergine versus Placebo: Global Rating (% improved), Outcome 5 Treatment length (weeks).

1.6. Analysis.

Comparison 1 Hydergine versus Placebo: Global Rating (% improved), Outcome 6 Mean Daily Dose.

1.7. Analysis.

Comparison 1 Hydergine versus Placebo: Global Rating (% improved), Outcome 7 Publication year.

Comparison 2. Hydergine versus Placebo: Comprehensive Rating.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Patient status	8	533	Mean Difference (IV, Fixed, 95% CI)	0.93 [0.52, 1.35]
1.1 Inpatient	3	177	Mean Difference (IV, Fixed, 95% CI)	1.46 [0.34, 2.57]
1.2 Outpatient	5	356	Mean Difference (IV, Fixed, 95% CI)	0.85 [0.40, 1.29]
2 Diagnosis	9	552	Mean Difference (IV, Fixed, 95% CI)	0.96 [0.54, 1.37]
2.1 Alzheimer's Disease	2	94	Mean Difference (IV, Fixed, 95% CI)	3.04 [‐0.39, 6.47]
2.2 Vascular Dementia	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Primary Dementia	4	259	Mean Difference (IV, Fixed, 95% CI)	2.16 [0.58, 3.75]
2.4 Cerebral Deterioration	3	199	Mean Difference (IV, Fixed, 95% CI)	0.83 [0.40, 1.27]
2.5 Cerebral Insufficiency	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Mean Age	6	486	Mean Difference (IV, Fixed, 95% CI)	0.88 [0.46, 1.30]
3.1 65‐75 Years	3	287	Mean Difference (IV, Fixed, 95% CI)	1.48 [‐0.05, 3.01]
3.2 75‐85 Years	3	199	Mean Difference (IV, Fixed, 95% CI)	0.83 [0.40, 1.27]
4 Gender	8	533	Mean Difference (IV, Fixed, 95% CI)	0.93 [0.52, 1.35]
4.1 0‐50% Male	6	497	Mean Difference (IV, Fixed, 95% CI)	0.90 [0.48, 1.32]
4.2 51‐100% Male	2	36	Mean Difference (IV, Fixed, 95% CI)	2.66 [‐0.41, 5.74]
5 Treatment length (weeks)	9	552	Mean Difference (IV, Fixed, 95% CI)	0.96 [0.54, 1.37]
5.1 Short (< 10 weeks)	1	19	Mean Difference (IV, Fixed, 95% CI)	11.60 [3.13, 20.07]
5.2 Medium (12‐16 Weeks)	5	369	Mean Difference (IV, Fixed, 95% CI)	2.25 [0.93, 3.58]
5.3 Long (> 23 weeks)	3	164	Mean Difference (IV, Fixed, 95% CI)	0.79 [0.35, 1.22]
6 Mean Daily Dose	9	552	Mean Difference (IV, Fixed, 95% CI)	0.96 [0.54, 1.37]
6.1 Low (1.5‐3.0mg/d)	2	94	Mean Difference (IV, Fixed, 95% CI)	2.52 [‐0.37, 5.42]
6.2 High (4.0‐9.0mg/d)	7	458	Mean Difference (IV, Fixed, 95% CI)	0.93 [0.51, 1.34]
7 Publication year	9	552	Mean Difference (IV, Fixed, 95% CI)	0.96 [0.54, 1.37]
7.1 Early (1955‐1973)	1	52	Mean Difference (IV, Fixed, 95% CI)	2.36 [‐0.02, 4.74]
7.2 Late (1974‐1990)	8	500	Mean Difference (IV, Fixed, 95% CI)	0.91 [0.49, 1.34]

2.1. Analysis.

Comparison 2 Hydergine versus Placebo: Comprehensive Rating, Outcome 1 Patient status.

2.2. Analysis.

Comparison 2 Hydergine versus Placebo: Comprehensive Rating, Outcome 2 Diagnosis.

2.3. Analysis.

Comparison 2 Hydergine versus Placebo: Comprehensive Rating, Outcome 3 Mean Age.

2.4. Analysis.

Comparison 2 Hydergine versus Placebo: Comprehensive Rating, Outcome 4 Gender.

2.5. Analysis.

Comparison 2 Hydergine versus Placebo: Comprehensive Rating, Outcome 5 Treatment length (weeks).

2.6. Analysis.

Comparison 2 Hydergine versus Placebo: Comprehensive Rating, Outcome 6 Mean Daily Dose.

2.7. Analysis.

Comparison 2 Hydergine versus Placebo: Comprehensive Rating, Outcome 7 Publication year.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Arrigo 1973.

Methods	Randomized Double‐blind Placebo‐controlled Parallel‐group Duration: 12 weeks
Participants	Country: Italy No. of centers: 1 Diagnosis: cerebrovascular insufficiency Diagnosis defined by: 7‐point rating scale Total No. of patients: 20 Setting: not described Age: Hydergine mean age = 66.1 (range = 59‐75), Placebo mean age = 70.5 (range = 63‐79) Gender: 19 males, 1 female Exclusions: not described
Interventions	Route: oral Treatment: Hydergine 1.5 mg t.i.d. Control: Placebo t.i.d.
Outcomes	Rating scale (1‐7 points) assessing physical parameters, subjective complaints, Activities of Daily Living, cognitive and emotional functioning, EEG, global rating
Notes	No. excluded after randomization: Total=0, Treatment=0, Control=0. No. not included in analysis: Total=0, Treatment=0, Control=0.

Chierichetti 1985.

Methods
Participants
Interventions
Outcomes
Notes

Cox 1978.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 12 weeks
Participants	Country: England No. of centers: 1 Diagnosis: varying degrees of dementia Diagnosis defined by: SCAG (Sandoz Clinical Assessment Geriatric Scale) Total No. of patients: 37 Setting: inpatient Age: Hydergine (range= 62‐93 years), Placebo (range= 68‐89 years) Gender: 5 males, 32 females Inclusion: Sandoz Clinical Assessment Geriatric > 2 points in 6 symptoms areas, and > 3 points in more than 2 designated symptoms. Exclusion: psychosis, post‐traumatic brain disease, post‐infective brain disease, cerebral neoplasm, marked mental deterioration, psychotropic drugs, hypotensive drugs containing reserpine or with severe intercurrent illness.
Interventions	Route: oral Treatment: Hydergine 1.5 mg t.i.d. Control: Placebo with multivitamin tablets and 600 mg of potassium t.i.d.
Outcomes	SCAG (Sandoz Clinical Assessment Geriatric), other mental functioning tests, global rating
Notes	No. excluded after randomization: Total=9, Treatment=5, Control=4. No. not included in analysis at 6 weeks: Total=7, Treatment=3, Control=4. No. not included in analysis at 12 weeks: Total=9, Treatment=5, Control=4.

Exton‐Smith 1982.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 9 weeks
Participants	Country: England No. of centers: 1 Diagnosis: multi‐infarct dementia with vascular component, or senile dementia Alzheimer's type Diagnosis defined by: Crichton Royal Rating Scale for SDAT and HIS (Hachinski Scale) >6 for MID Total No. of patients: 20 Setting: not described Age: range= 68‐94 Gender: not described Exclusion: medications such as vasodilators, CNS stimulants, major tranquilizers, antidepressants, daytime sedatives, psychiatric illness other than dementia, physical disorders which might change and affect assessment of mental status or ability to complete the study.
Interventions	Route: oral Treatment: Hydergine 3mg t.i.d. Control: Placebo t.i.d.
Outcomes	Automated Picture Matching Task Raven's Progressive Matrices Word recall (immediate and delayed) Orientation (general) activities of daily living The MACC Behavioural Adjustment Scale
Notes	No. excluded after randomization: Total=1. Treatment=1 Control=0. No. not included in analysis: Total=1. Treatment=1, Control=0.

Gaitz 1977.

Methods	RandomizedDouble‐blindPlacebo‐controlledParallel‐groupDuration: 6 months
Participants	Country: United StatesNo. of centers: 1Diagnosis: Nursing home residents with organic brain syndrome associated with senile dementia, cerebral arteriosclerosis, or both.Diagnosis defined by: SCAG (Sandoz Clinical Assessment Geriatric Scale)Total No. of patients: 54 (Hydergine = 28, Placebo = 26)Setting: nursing homeAge: > 59 years of ageGender: not describedInclusion: Sandoz Clinical Assessment Geriatric (SCAG) score > 4 on at least 6 items and a score > 4 on 2 of 7 symptoms.Exclusion: significant preexisting or concomitant "functional" psychosis (schizophrenia, major affective disorders), marked mental retardation, posttraumatic or postinfective brain damage, cerebral neoplasm, and any other organic disorder not related to senile cortical atrophy or cerebrovascular disease.
Interventions	Route: sublingualTreatment: Hydergine 1 mg t.i.d. (2 tablets [.5mg each] 30 minutes before meals) Control: 2 Placebo tablets t.i.d. (30 minutes before meals)
Outcomes	SCAG, 7‐point rating scale, vital signs, labs, physical examinations, MSCL (Mental Status Check List), global rating
Notes	No. excluded after randomization: Total=7, Treatment=5, Control=2.No. not included in analysis: Total:7, Treatment=5, Control=2.Note: Total of 19 subjects (Treatment=12, Control=7) were not observed for full 24 weeks (only 12 weeks) but used in analysis because of data substitution rule if in trial >49% of time.

Giove 1973.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 6 months
Participants	Country: Italy No. of centers: 1 Diagnosis: cerebrovascular insufficiency Diagnosis defined by: 7‐point rating scale Total No. of patients: 30 Setting: Inpatient Age: Mean= 79 Gender: not described Exclusions: not described
Interventions	Route: oral Treatment: Hydergine 1.5 mg t.i.d. Control: Placebo t.i.d.
Outcomes	Rating scale (1‐7 points) assessing physical parameters, subjective complaints, activities of daily living, cognitive and emotional functioning, EEG, global rating
Notes	No. excluded after randomization: not described No. not included in analysis: not described

Hollister 1955.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 60 days
Participants	Country: United States No. of centers: 1 Diagnosis: Hospitalized patients with various diagnosis (chronic brain syndrome, cerebral arteriosclerosis, pre‐senile brain disease, arterial hypertension, involutional psychotic reaction, schizophrenic reaction, syphillis, Korsakoff). Diagnosis defined by: "careful clinical assessment" Total No. of patients: 28 Setting: Inpatient Age: Median age= 66 (range= 56‐83) Gender: 27 males, 1 female Inclusion: hospitalized patients with various diagnosis (chronic brain syndrome, cerebral arteriosclerosis, pre‐senile brain disease, arterial hypertension, involutional psychotic reaction, schizophrenic reaction, syphillis, Korsakoff). Exclusion: unclear
Interventions	Route: sublingual Treatment: Hydergine 30d= 1mg q.i.d., 15d= 1mg t.i.d., 15d= 1mg b.i.d. Control: Placebo 30d= q.i.d., 15d= t.i.d., 15d= b.i.d.
Outcomes	Hospital Adjustment Scale, physical examination, blood pressure, global rating
Notes	No. excluded after randomization: Total = 2, Treatment = unclear, Control = unclear. No. not included in analysis: Total = 2, Treatment = unclear, Control = unclear.

Lazzari 1983.

Methods	RandomizedDouble‐blindPlacebo‐controlledParallel‐groupDuration: 12 months
Participants	Country: Italy No. of centers: 40 Diagnosis: chronic senile cerebral insufficiency Diagnosis defined by: clinical examination Total no. of patients: 458 Age: Hydergine mean= 75.8 (SD=6.9), Placebo= 76.0 (SD=7.9) Gender: Hydergine: 73 males, 122 females, Placebo: 82 males, 111 females Inclusion: Scores > 2 on Sandoz Clinical Assessment Geriatric (SCAG) items (confusion, attention, short term memory, disorientation, anxiety, depression, emotional lability), Hachinski score between 5‐14. Exclusion: Severe hepatic, renal or cardiac disease, psychosis, alcoholism, pregancy, vascular accident, recent trauma that would affect stability, physical problems affecting testing (e.g., tremor, involuntary movements); psychotropic or CNS medications within 4 months of the study (e.g., antihypertensives, beta‐blockers).
Interventions	Route: not describedTreatment: Hydergine 1.5mg t.i.d.Control: Placebo t.i.d.
Outcomes	SCAG, Geriatric Rating Scales, Northwestern University Disability Scales, Nowlis test, Digit Symbol, Digit Span, Toulouse‐Pieron test
Notes	No. excluded after randomization and excluded from analysis: Total=184. Treatment=84, Control=100.

McConnachie 1973.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 12 weeks
Participants	Country: England No. of centers: 2 Diagnosis: mild‐moderate symptoms of cerebrovascular insufficiency or senile dementia Diagnosis defined by: Crichton Royal Behavioral Rating Scale (RBRS) Total No. of patients: 58 Age: Hydergine mean= 81, Placebo mean= 82 Gender: 10males, 42 females Inclusion: hospital inpatients; nursing home residents; score 2 or 3 on 7/10 factors on Crichton RBRS with total score of 22‐28. Exclusion: psychoses; post‐traumatic brain damage; regular use of psychotropic or hypotensive medications.
Interventions	Route: oral Treatment: Hydergine 1.5mg t.i.d. Control: Placebo t.i.d.
Outcomes	Physical examination, Physician assessment ‐ 3 point scale on four main symptom complexes: 1) Physical manifestations (anorexia, headache, vertigo, tremor, muscle cramps, paraesthesia, fatigue), 2) Daily living activities (walking, eating, washing, dressing, general self care), 3) Attitudes and mood (hostility, emotional lability, unsociability, depression, uncooperativeness, indifference to surroundings, lack of motivation, anxiety, fear), 4) Motor activity (disinclination for activity or restlessness), global rating
Notes	No. excluded after randomization: Total = 6, Treatment = 4, Control = 2. No. not included in analysis: Total = 6, Treatment = 4, Control = 2.

McDonald 1985a.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 24 weeks (12 weeks, option for additional 12 weeks)
Participants	Country: United States No. of centers: 1 Diagnosis: primary degenerative dementia Diagnosis defined by: DSM‐III Duration of disease: Treatment=3.9 years (S.D.=1.66, range2‐7), Placebo=2.7 years (SD=1.42, range 1‐5). No. of patients: 20 Age: Hydergine mean: 64.6 (SD=6.1), Control: 65.1 (SD=8.9) Sex: 19 men, 1 woman Inclusion: normal ECG, vital signs, clinical labs, and GDS range (2‐4) Exclusions: physical abnormalities that would interfere with study, neurological or psychiatric diagnosis other than primary degenerative dementia.
Interventions	Route: oral Treatment: Hydergine 2 mg t.i.d. Control: Placebo t.i.d.
Outcomes	GDS (Global Deterioration Scale), MMSE (Mini‐Mental Status Exam) HDRS (Hamilton Depression Rating Scale), SCAG (Sandoz Clinical Assessment Geriatric Scale), IPSC‐E (The Inventory of Psychic and Somatic Complaints of the Elderly), Self‐Rating Scale, GDR (The Geriatric Depression Rating Scale) DAT (Divided Attention Task), SRT (Buschke Selective Reminding Test) Digit Symbol (WAIS), Trail‐Making Test
Notes	Initial 4 week single‐blind, placebo‐controlled design (prior to eligibility). No. excluded after randomization: Total=0. Treatment=0, Control=0. No. not included in analysis: (12 weeks) Total=0. Treatment=0, Control=0. No. not included in analysis : (24 weeks) Total=3. Treatment=1, Control=2.

McDonald 1985b.

Methods	Randomized Double‐blind Placebo‐controlled Parallel‐group Duration: 12 weeks
Participants	Country: United States No. of centers: 5 Diagnosis: primary degenerative dementia Diagnosis defined by: DSM‐III Total No. of patients: 236 Age: Hydergine mean age was 69.8, range (55‐79), Placebo mean age was 68.5, range (55‐79) Sex: Hydergine: 59 females, 43 males, Placebo: 46 females, 47 males. Inclusion: Global deterioration score 2‐4; Hamilton Depression Rating Scale (HDRS) <17; normal pre‐treatment ECG, vitals, labs; no additional neurological or psychiatric diagnosis. Exclusion: Significant abnormal findings that would interfere with study objectives.
Interventions	Route: oral Treatment: Hydergine 2mg t.i.d. Control: Placebo t.i.d.
Outcomes	GDS (Global Deterioration Scale), MMS (Mini Mental Status), HRDS (Hamilton), (Depression Rating Scale), SCAG (Sandoz Clinical Assessment Geriatric), IPSC‐E (Inventory of Psychic and Somatic Complaints), GDR (Geriatric Depression Rating Scale), DSST (Digit Symbol Substitution Test: subtest of WAIS: Wechsler) Trail‐Making Test, BSRT (Buschke Selective Reminding Test), physical examination labs, vitals, ECG, adverse reactions
Notes	Initial 2 week single‐blind, placebo controlled design (prior to elegibility). No. excluded after randomization: Total = 41. Treatment = 17, Control = 24. No. not included in analysis: Total: 41. Treatment = 17, Control = 24.

Peltz 1969.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 16 weeks
Participants	Country: United States No. of centers: 1 Diagnosis: Symptoms attributable to cerebral arteriosclerosis or simple senile degeneration Diagnosis defined by: Munroe III Visual Test Total No. of patients: 48 Age: Mean age was 82.5, range (64‐96), Hydergine mean age was 81.7, range (72‐95), Placebo mean age was 84, range (64‐96). Sex: 20 females, 13 males Inclusions: Hospitalized geriatric patients manifesting symptoms of cerebral arteriosclerosis or simple cerebral degeneration and exhibiting impairment in general daily living skills and social adaptability. Exclusions: schizophrenia, post‐traumatic brain damage, post‐infective brain disease, cerebral neoplasm; severe degeneration causing inability to participate in testing; psychotropic or sedative medications within 3 weeks prior to study.
Interventions	Route: sublingual Treatment: Hydergine .5mg t.i.d. Control: Placebo t.i.d.
Outcomes	Munroe III Visual Test, Physical examinations, labs, vital signs, Daily Living and Behavior Rating Scale, Geriatric Daily Living Rating Scale, Overall Rating of Severity (of degeneration), Global Rating of Improvement or Regression
Notes	No. excluded after randomization: Total = 15. Treatment = unclear, Control = unclear. No. not included in analysis: Total = 15. Treatment = unclear, Control = unclear.n

Rodriguez 1971.

Methods	RandomizedDouble‐blindPlacebo‐controlledParallel‐groupDuration: 12 weeks
Participants	Country: United StatesNo. of centers: 1Diagnosis: chronic cerebrovascular insufficiency associated with arteriosclerosisDiagnosis defined by: > 1 of the following (confusion, impairment of recent memory, impairment of mental alertness, deviations from normal mood)Total No. of patients: 60Age: Hydergine mean age was 71.9, range (62‐92), Placebo mean age was 73.0, range (62‐91)Sex: Hydergine: 15 female, 15 male, Placebo: 15 female, 15 maleInclusion: > 1 of the following (confusion, impairment of recent memory, impairment of mental alertness, deviations from normal mood)Exclusion: (1) psychosis, post‐traumatic brain damage, post infective brain disease, cerebral neoplasm, marked mental deterioration; (2) regular use of psychotropic drugs, sedatives, vasodilators; (3) physical conditions of such severity as to preclude participation; (4) unstable chronic physical illness; (4) instability in hospital setting; (5) expectation of discharge in < 12 weeks
Interventions	Route: sublingualTreatment: Hydergine 2, .5 mg tablets t.i.d.Control: Placebo 2, .5mg tablets t.i.d.
Outcomes	Sandoz Clinical Assessment Geriatric, Clinical Global Impression
Notes	No. excluded after randomization and not included in analyses: Total=1. Treatment=1, Control=0.

Rouy 1989.

Methods	Randomized Double‐blind Placebo‐controlled Parallel‐group Duration: 6 months
Participants	Country: France No. of centers: 1 Diagnosis: mild‐moderate mental deterioration due to cerebral aging Diagnosis defined by: SCAG (Sandoz Clinical Assessment Geriatric Scale) Total No. of patients: 97 Age: Hydergine mean age was 82.9, range (63‐98), Placebo mean age was 82.6, range (71‐95) Sex: Hydergine: 34 female, 13 male, Placebo: 32 female, 18 male Inclusion: geriatric in‐patients diagnosed with mild‐moderate mental deterioration due to cerebral aging and SCAG score of 3‐5 on 11 items. Exclusion: vascular accident < 3mth prior to study; cardiac arrhythmias; carotid souffle; severe psychosis; organic brain lesion; metabolic, toxic, and infectious encephalopathy; severe mental shock or physical disease inducing psychic effects; no medications affecting metabolism, cerebral or peripheral vascular system.
Interventions	Route: oral Treatment: Hydergine 4.5mg once‐a‐day Control: Placebo 1 tablet per day
Outcomes	EACG ( a French version of SCAG), NOSIE (Nurse's Observation Scale for In‐Patients) Record of adverse effects, Global Rating
Notes	No. excluded after randomization: Total = 10. Treatment = 5, Control = 5. No. not included in analysis: Total = 10. Treatment = 5, Control = 5.

Short 1972.

Methods	Randomized Double‐blind Placebo‐controlled Parallel‐group Duration: 12 weeks
Participants	Country: United States No. of centers: 1 Diagnosis: general symptoms of aging attributed to chronic degenerative processes such as cerebrovascular insufficiency associated with arteriosclerosis. Diagnosis defined by: unclear Total No. of patients: 60 Age: Hydergine mean age was 85, range (68‐96), Placebo mean age was 81, range (71‐93). Sex: Hydergine: 21 females, 2 males, Placebo: 22 females, 4 males. Inclusion: hospitalized geriatric patients manifesting general symptoms of aging such as deficits in mental alertness, recent memory, mood (depression), and thinking (confusion). Exclusion: severe mental disorders affecting treatment; upcoming traumatic events; psychoactive or vasodilator medications three weeks prior to study.
Interventions	Route: sublingual Treatment: Hydergine 2.5mg t.i.d. Control: Placebo t.i.d.
Outcomes	Clinical status (17 symptoms, 7‐point scale), Overall clinical status, Global therapeutic response, Mental status, labs, vitals, physical examinations, untoward effects (5 symptoms, 7‐point scale)
Notes	No. excluded after randomization: Total = 11. Treatment = 7, Control = 4. No. not included in analysis: Total = 11. Treatment = 7, Control = 4.

Thompson 1990.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 24 weeks
Participants	Country: United States No. of Centers: 1 Diagnosis: Probable Alzheimer's disease Diagnosis defined by: DSM‐III (Diagnostic and Statistical Manual of Mental Disorders‐ Third Edition), NINCDS‐ADRDA (National Institute of Neurological and Communicative Disorders and Strokes, and the Alzheimer's Disease and Related Disorders Association). Inclusion: Modified Hachinski Ischemic Scoring Scale < 4. Mini Mental State Examination: range (10‐23) Hamilton Psychiatric Rating Scale for Depression <19 Inventory of Psychic and Somatic complaints in the Elderly Exclusions: evidence of multi‐infact on CT, MRI, labs, or physical exam Total No. of patients: 80 Sex: 46 women, 34 men Age: Hydergine mean was 72, Placebo mean was 70, range (55‐79)
Interventions	Route: oral Treatment: Hydergine‐LC 1mg t.i.d. Control: Placebo t.i.d.
Outcomes	WAIS (Wechsler Adult Intelligence Scale), Digit Symbol Substitution Test, Russell revision of the Wechsler Memory Scale (Logical Memory and Visual Reproduction), IPSC‐E (Inventory of Psychic and Somatic Complaints in the Elderly), SCAG (Sandoz Clinical Assessment Geriatric Scale), GERRI (Geriatric Evaluation by Relatives Rating Instrument), HDRS (Hamilton Psychiatric Rating Scale for Depression)
Notes	No. excluded after randomization: not stated No. not included in analysis: Total=12, Treatment=6, Control=6.

van Loveren 1984.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 24 weeks
Participants	Country: Netherlands No. of centers: 3 Diagnosis: mild‐moderate senile mental deterioration Diagnosis defined by: SCAG (Sandoz Clinical Assessment Geriatric Scale) Total No. of patients: 58 Setting: Old‐peoples homes Age: Hydergine mean age = 83.7, Placebo mean age = 82.7 Sex: unclear Inclusion: geriatric residents diagnosed with mild‐moderate senile mental deterioration, including SCAG scores of 3‐5 on >5 symptoms and 4‐5 on >2 key symptoms and HIS (Hachinski Ischemic Score) of <7. Exclusion: HIS >7; psychiatric, neurologic, or other severe systemic disorders; vasodilator or CNS medications; SCAG score of 7 on 1 or more items.
Interventions	Route: oral Treatment: Hydergine 1.5mg t.i.d. (1st 12 wks) and Hydergine 1.5mg (AM) and 3mg b.i.d. (Noon and PM) for 2nd 12wks. Control: Placebo t.i.d.
Outcomes	SCAG Behavior, assessed by 45 min. conversation Hamburg Wechsler Adult Intelligence Scale subtests: Digit Span Digit Symbol Block Design Labyrinth Test Benton Visual Retention Trails A & B Questionnaires on biographical data, activities of daily living, personality, and activities Physical examination, labs, vitals, ECG
Notes	No. excluded after randomization: Total = 8. Treatment = 2, Control = 6. No. not included in analysis: Total = 8. Treatment = 2, Control = 8.

Wilder 1973.

Methods	Randomized Double‐blind Placebo‐controlled Parallel‐group Duration: 6 weeks
Participants	Country: United States No. of centers: 1 Diagnosis: cerebrovascular insufficiency associated with thrombo‐embolic stroke Diagnosis defined by: physician assessment and hospital record Total No. of patients: 28 Age: Hydergine mean age was 68, range (55‐90), Placebo mean age was 68, range (56‐79). Sex: 28 male Inclusion: Hospitalized geriatric patients who had thrombo‐embolic stroke resulting in fixed neurological deficits 4‐15 weeks prior to study; mild‐moderate in >1 of 5 CVI (cerebrovascular insufficiency) symptoms. Exclusion: unclear
Interventions	Route: sublingual Treatment: Hydergine 2 tablets .5mg t.i.d. Control: Placebo t.i.d.
Outcomes	SCAG (Sandoz Clinical Assessment Geriatric Scale), Mental Status Exam, Neurologic Status (Cortical Function, Cranial Nerves, Reflexes, Sensation, Miscellaneous, Global Change), EEG, EKG, labs, vitals, physical examination, Global Rating
Notes	No. excluded after randomization: Total = 0. Treatment = 0, Control = 0. No. not included in analysis: Total = 0. Treatment = 0, Control = 0.

Winslow 1972.

Methods	Randomized Double‐blind Parallel‐group Placebo‐controlled Duration: 12 weeks
Participants	Country: United States No. of centers: 1 Diagnosis: chronic cerebrovascular insufficiency due to cerebral arteriosclerosis Diagnosis defined by: Assessment of Clinical Status Total No. of patients: 59 Age: range (60‐94), Hydergine mean was 77.9, range (60‐94); Placebo mean age was 76.3, range (62‐89). Sex: Hydergine: 15 females, 10 males; Placebo: 12 females, 13 males Inclusion: hospitalized geriatric patients suffering from symptoms attributable to diagnosis of chronic cerebrovascular insufficiency due to cerebral arteriosclerosis Exclusion: psychosis, post‐traumatic brain damage, post‐infective brain disease, cerebral neoplasm, mental deterioration; regular use of psychotropic, sedatives or vasodilator medications 3 weeks prior to study; insufficient symptoms for diagnosis; physical conditions precluding participation; acute physical disorders affecting assessment; expectation of transfer, traumatic events, or discharge before end of study.
Interventions	Route: sublingual Treatment: Hydergine 1mg t.i.d. Control: Placebo t.i.d.
Outcomes	Assessment of Clinical Status (19 items, 7‐point rating scale), vitals, labs, physical examination, Drug Reaction Record, Global Rating of Change, Major Symptom Inventory
Notes	No. excluded after randomization: Total = 9. Treatment = 4, Control = 5. No. not included in analysis: Total = 9. Treatment = 4, Control = 5.

Yesavage 1981.

Methods	Randomized Double‐blind Parallel group Placebo‐controlled Duration: 12 weeks
Participants	Country: United States No. of centers 1 Diagnosis: mild to moderate senile dementia Total number of patients: 19 Sex: 19 males Age: greater than 54 years Inclusion: SCAG (Sandoz Clinical Assessment Geriatric Scale) >4 points for 6 target symptoms of affect state and cognitve function; at least >5 symptoms had to be related to cognitive function. SCAG overall impression rating of 4‐5 indicating mild to moderate degree of senile dementia Exclusions: Parkinson's disease, Huntington's chorea, affective, toxic, organic, or presile psychosis, schizophrenia, posttraumatic encephalopathy, delirium, drug induced brain changes, syphillis, cerebral neoplasm, normal pressure hydrocephalus, mental retardation, alcoholic brain syndrome, cerebral vascular accident with residual neurological sequelae, acute brain disease associated with systematic disease, or endocrine, metabolic, or hematologic disorders; recent history of long‐term psychoactive chemotherapy, known hypersensitivity to DEM, blindness, deafness, language difficulties, current alcohol or hallucinogen abuse, administration of an investigational drug 4 weeks immediately preceding admission to study, administration of any drug known to cause major organ system toxicity during three months preceding participation in current study.
Interventions	Route: oral Hydergine 1 mg t.i.d. Control: Placebo t.i.d.
Outcomes	SCAG (Sandoz Clinical Assessment Geriatric Scale), Raskin Depression Scale, Buschke Word Recall Test
Notes	No. excluded after randomization: Total=0, Treatment=0, Control=0. No. not included in analysis: Total=0; Treatment=0, Control=0.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Ammon 1995	Review article
Anon 1970	letter to the editor with insufficient information
Anon 1990	not a clinical trial; a critique
Arrigo 1969	outcome was not reported adequately
Arrigo 1972	outcome measure was irrelevant to clinical efficacy
Arrigo 1975	outcome was presented in other reports
Arrigo 1989	intravenous administration of Hydergine
Atarashi 1988	trial was not placebo‐controlled
Baldoni 1971	trial was not placebo‐controlled
Balestreri 1984	trial was not placebo‐controlled
Banen 1972	outcome was not reported adequately
Bargheon 1973	outcome not adequately reported
Bastos 1980	trial was not placebo‐controlled
Battaglia 1990	trial was not placebo controlled
Bazo 1972	trial was not placebo‐controlled
Bazo 1973	trial was not placebo‐controlled
Bente 1979	outcome measure was irrelevant to clinical efficacy
Benton 1951	patient sample was not appropriate for medication (sample not homogeneous)
Biel 1976	cross‐over trial without first‐period data
Bochner 1973	patient sample was not appropriate for medication (stroke patients ‐motor functioning assessment)
Bohard 1967	patient sample was not appropriate for medication (psychotic patients)
Brage 1984	data reported do not permit an empirical analysis of outcome
Breeze 1985	outcome measure was irrelevant to clinical efficacy
Breeze 1988	outcome measure was irrelevant to clinical efficacy
Cai 2003	No placebo group
Casale 1989	intravenous administration of Hydergine
Chen 1997	No placebo control
Cheng 1998	no placebo control group
Chien 1980	patient sample was not appropriate for medication (data not explicit; sample population = tardive dyskinesia)
Chudnovsky 1979	data reported do not permit an empirical analysis of outcome
Cid 1975	outcome measure was irrelevant to clinical efficacy
Cook 1982	intravenous administration of Hydergine
Dartigues 1983	data reported do not permit an empirical analysis of outcome
de Brito‐Paiva 1982	data reported do not permit an empirical analysis of outcome
DeLaRevilla 1967	patient sample was not appropriate for medication (hypertension)
Dennler 1979	trial was not placebo‐controlled
Desimirovic 1988	trial was not placebo‐controlled
Ditch 1971	outcome was not reported adequately
Einspruch 1976	trial was not placebo‐controlled
Forster 1955	treatment duration was not described
Gentili 1976	data reported do not permit an empirical analysis of outcome
Gerin 1969	outcome was not adequately reported
Gerin 1970
Glover 1980	letter to the editor with insufficient information
Good 1982	data reported do not permit an empirical analysis of outcome
Grill 1969	outcome was not adequately reported
Grobe Einsler 1993	Review article
Gross 1979	outcome measure was irrelevant to clinical efficacy
Heiss 1971	data reported do not permit an empirical analysis of outcome
Herzfeld 1972a	data reported do not permit an empirical analysis of outcome
Herzfeld 1972b	outcome measure was irrelevant to clinical efficacy
Hofstatter 1956	data reported do not permit an empirical analysis of outcome
Hollingsworth 1974
Hollingsworth 1980	outcome was not adequately reported
Hoyer 1969	letter to the editor with insufficient information
Huang 2003	no placebo control
Huber 1986	patient sample was not appropriate for medication (sample has no disorder)
Hughes 1976	Review article
Irfan 1978a	outcome was not adequately reported
Jansen 1985	trial was not placebo controlled
Jarvik 1981	data reported do not permit an empirical analysis of outcome
Jenike 1986	cross‐over trial without first‐period data
Jenike 1989	Review article
Jenike 1990	No placebo control
Jennings 1972	outcome was not adequately reported
Junod 1978	data reported do not permit an empirical analysis of outcome
Kaiser 1956	open trial
Kanowski 1988	outcome was not adequately reported
Klein 1985	open trial
Koberle 1984
Kugler 1978	data reported do not permit an empirical analysis of outcome
Kugler 1985	trial was not placebo‐controlled
Kuskowski 1990
Labecki 1954	intravenous administration of Hydergine
Ladurner 1991	trial was not placebo controlled
Li 2004	no placebo control
Liao 2004	no placebo control
Linden 1975	data reported do not permit an empirical analysis of outcome
Linden 1976
Linder 1977
Liu 2005	no placebo control
Lozeron 1974	open trial
Luo 2001	No placebo control
Mars 1970	trial was not placebo‐controlled
Martucci 1986	outcome was not adequately reported
Matejcek 1979	data reported do not permit an empirical analysis of outcome
Matejcek 1980
Matejcek 1986	outcome measure was irrelevant to clinical efficacy
McConnachie 1978	data reported do not permit an empirical analysis of outcome
McDonald 1979	review article
McDonald 1989	outcome was not adequately reported
Mellini 1973	trial was not placebo‐controlled
Memin 1973	trial was not placebo‐controlled
Misra 1982	data reported do not permit an empirical analysis of outcome
Misurec 1978	data reported do not permit an empirical analysis of outcome
Moglia 1983	outcome was not adequately reported
Nelson 1973	trial was not plalcebo‐controlled
Nelson 1975	trial was not placebo‐controlled
Nicrosini 1976	trial was not placebo‐controlled
Novo 1978	outcome was not adequately reported
Odanische 1981	trial was not placebo‐controlled
Olivella 1981	trial was not placebo‐controlled
Orma 1956	treatment was not double‐blind
Oswald 1979	outcome was not adequately reported
Oswald 1980	trial was not placebo‐controlled
Oswald 1982	trial was not placebo‐controlled
Ouaniche 1981	trial was not placebo‐controlled
Parade 1978	trial was not placebo‐controlled
Patin 1981	outcome was not adequately reported
Paux 1975	outcome was not adequately reported
Pfeiff 1980	trial was not placebo‐controlled
Piguet 1981	trial was not placebo‐controlled
PiresDeOliveira 1973	outcome measure was irrelevant to clinical efficacy
Pomara 1983	trial was not placebo‐controlled
Popkin 1956	intravenous administration of Hydergine
Puxty 1989	data reported do not permit an empirical analysis of outcome
Puxty 1991
Pöpperl 2004	No placebo control
Rao 1972	outcome was not adequately reported
Rehman 1973a	data reported do not permit an empirical analysis of outcome
Rehman 1973b	data reported do not permit an empirical analysis of outcome
Riccardi 1978	trial was not placebo‐controlled
Ronge 1982	trial was not placebo‐controlled
Rosen 1975	trial was not placebo‐controlled
Roubicek 1971
Roubicek 1972	Data reported do not permit an empirical analysis of the outcome.
Roubicek 1973
Roubicek 1975
Saletu 1990	Sample were not demented
Saletu 1994	trial was not placebo‐controlled
Samorajski 1982	outcome measure was irrelevant to clinical efficacy
Schardt 1982	patient sample was not appropriate for medication (hypertension)
Schartl 1978	trial was not placebo‐controlled
Schneider 1994	review article
Schnell 1967	outcome measure was irrelevant to clinical efficacy
Setnikar 2001	Healthy participants
Seus 1969	data reported do not permit an empirical analysis of outcome
Shoptaw 2005	Cocaine‐dependent participants
Soni 1975	outcome was not adequately reported
Spiegel 1983
Spilich 1996	trial was not placebo‐controlled
Stewart 1996	trial was not placebo‐controlled
Strauss 1951	open trial
Strauss 1954	patient sample was not appropriate for medication (organic peripheral vascular disease)
Tartara 1974
Tecce 1981	outcome was presented in other reports
Tecce 1983a	outcome measure was irrelevant to clinical efficacy
Tecce 1983b	outcome was presented in other reports
Teixeira 1970	cross‐over trial without first‐period data
Thibault 1974	data reported do not permit an empirical analysis of outcome
Thienhaus 1987	data reported do not permit an empirical analysis of outcome
Tinklenberg 1986	outcomes not adequately reported
Triboletti 1969	outcome was not adequately reported
Tucker 1982	open trial
Winslow 1974	trial was not placebo‐controlled
Yan 2001	No placebo control
Yesavage 1979	trial was not placebo‐controlled
Yesavage 1981b	trial was not placebo‐controlled
Yoshikawa 1983	trial was not placebo‐controlled
Zhao 1999	No placebo control

Contributions of authors

‐Jason Olin assisted in article collection, data analysis, and wrote the primary draft of the manuscript. 
 ‐Lon Schneider co‐wrote and assisted with editing the manuscript, and assessed search results for the 2009 update. 
 ‐Adrian Novit and Susan Luczak selected articles, coded article content, input article related data into Revman, and edited the manuscript.

Sources of support

Internal sources

• No sources of support supplied

External sources

• National Institute of Mental Health (MH01368, MH19074, MH48759), USA.

• Sandoz Pharmaceuticals, Inc., Switzerland.

• National Institute on Aging (AG05142), USA.

Declarations of interest

The authors (JO, LS) were involved in a previous meta‐analysis of this literature that was commissioned by Sandoz Pharmaceuticals Corporation (now Novartis).

New search for studies and content updated (no change to conclusions)