Prescribing patterns in dementia: a multicentre observational study in a German network of CAM physicians

Physicians

Of the 22 physicians, 17 were GPs (77%) and 5 were specialist GPs (23%). The participating physicians did not differ significantly from the overall population of physicians certified in anthroposophy in Germany (n = 362) in terms of age (mean = 49.4; SD = 6.3 years vs. mean = 47.5; SD = 6.1 years; P = 0.709) or gender (60.0% vs. 62.2% men; P = 0.917), and were only slightly younger and consisted of a similar percentage of women compared to all office-based physicians in Germany (mean 52.0 years; 61.2% men) [31].

Study population

Anti-dementia drug therapy

56.7% of patients with dementia with Alzheimer's disease received anti-dementia drugs, whereas less than 25% of patients with other dementia types were prescribed anti-dementia drugs (Table 3). Donepezil and memantine were prescribed mostly for patients with dementia with Alzheimer's disease. The phytotherapeutical Ginkgo biloba was prescribed for patients with dementia with Alzheimer's disease, vascular dementia, and unspecified dementia in equal shares (16% - 20%). Patients with vascular dementia were prescribed Ginkgo biloba (15.9% Ginkgo biloba vs. 2.0% of any other anti-dementia drug) nearly exclusively.

In total, 448 prescriptions of anti-dementia drugs for the 148 patients who received at least one of these drugs were reported. The phytopharmaceutical Ginkgo biloba was the most frequently prescribed anti-dementia drug overall (67.6% of all anti-dementia drugs).

Prescriptions for dementia-related co-morbidities

There were significant differences in the proportion of patients treated with an antidepressant (GP: 17.6%; neurologist: 28.1%; P = 0.003) and benzodiazepine (GP: 11.4%; neurologist: 6.4%; P = 0.039) according to physician's specialisation. No difference was found for the treatment with antipsychotics (GP: 15.8%; neurologist: 14.3%; P = 0.635).

Factors related to any anti-dementia drug therapy

Factors related to the prescription of Ginkgo biloba

The prescription of Ginkgo biloba was associated with sex, patient age, physician's specialisation, and the type of dementia (Table 5). Older patients (AOR = 1.06; CI: 1.02-1.10) were more likely to receive Ginkgo biloba. Female sex (AOR = 0.41; CI: 0.19-0.89), treatment by a neurologist (AOR = 0.09; CI: 0.03-0.23), and the diagnosis of Alzheimer's disease (AOR = 0.07; CI: 0.04-0.16) rapidly reduced the odds of being prescribed Ginkgo biloba, whereas co-morbidity, co-medication, consultation type, and the year of prescription had no influence.

Study population

The median age of our study population was 81 years. Compared to the recently published study of Olsson et al. [32] conducted in nursing homes and special care units, our patient population was slightly younger due to the outpatient-character of this study but comparable with data from German health insurance provided by Koller et al. [33]. 69% of our patients were female, which is similar to the proportion of women in dementia patients in Germany (70% female) [34]. Considering the high age of our population the frequent number of co-morbidities is not surprising. Hypertension, congestive heart failure, and ischemic heart diseases are particularly common in this age and are well known as primers for vascular dementia and thus their occurrence is an indicator for dementia in our population (see Table 2).

Anti-dementia drug therapy

In our study almost 26% of patients received anti-dementia drug therapy. According to the current ATC classification, Ginkgo biloba is listed in the group of anti-dementia drugs together with cholinesterase inhibitors and memantine. Thus we counted Ginkgo biloba as an anti-dementia drug although the literature is not consistent in this case. This has to be taken into consideration when comparing our results with other studies.

According to Formiga et al., 80.6% of the patients in a Spanish population received cholinesterase inhibitors or/and memantine for Alzheimer's disease [35]. In an Italian survey from 2007, Frisoni found a prescription rate of cholinesterase inhibitors in 90% of patients with Alzheimer's disease and 35-45% with vascular dementia. Other drugs, such as Ginkgo biloba, were prescribed less frequently, except for vascular dementia (20%) [36]. For Germany, Stoppe et al. found that only 24% of patients with Alzheimer's disease were prescribed cholinesterase inhibitors in 2004 [37], which is comparable to our study.

Although both the results of our study and those of Stoppe et al. [37] might be surprising at first, it has to be taken into account that especially the DEGAM guideline recommends a cautious and responsible handling of anti-dementia drugs in dementia patients and thus the results are in line with both the guidelines and the experience of the family physicians [25]. In Parkinson's disease, only one of 42 patients was mainly treated with rivastigmine which is recommended in the neurological guideline [19].

In 2006, a total of 50% of all prescribed anti-dementia drugs in Germany were cholinesterase inhibitors and memantine [13]. The most often prescribed anti-dementia drug in Germany was the cholinesterase inhibitor donepezil with 37% [38], which with a rate of 45% also plays a predominate role in our study. Traditional anti-dementia drugs like ergot alkaloids and piracetam, which are not recommended in the treatment guidelines, have declined in their prescribing rates over in recent years and, with only 6 patients, also play a minor role in our setting.

In our study only 8 out of 148 patients (5.4%) were prescribed both cholinesterase inhibitors and memantines during the study period. This is comparable to Hoffmann et al. who found only one anti-dementia drug in over 90% of his patients and Truter who only reported 5.3% of patients that were prescribed more than one active ingredient for Alzheimer's disease during one year (mostly donepezil or galantamine, and memantine) [38, 39], whereas Formiga found combined therapy of cholinesterase inhibitors in 15.5% of patients [35]. While the DEGAM guideline disapproves combination therapies due to their low level of evidence, the DGN/DGPPN guideline refers to 2 RCTs and claims that an add-on therapy might be considered [18, 19]. However, it has to be noted that in our study concomitant treatment with two drugs may also be a result of the course of therapy and might not reflect a factual combination therapy.

Factors related to any anti-dementia drug therapy

According to Formiga, lower Barthel Index scores predict fewer prescriptions of pharmcological therapy in Alzheimer's disease [40]. Other factors like patient age, gender, co-morbidity, neuroleptic and antidepressant therapy, and hospital admission in the previous year show no significance after adjusting the multivariate logistic regression model.

We found that type of dementia, physician specialisation, co-morbidity, and neuroleptic therapy influenced the prescription of any anti-dementia drug (i.e. cholinesterase inhibitors, memantine and/or Ginkgo biloba). While odds ratios for Alzheimer's disease are presumably higher than those for vascular dementia, it is in accordance with the results of Formiga and Frisoni [36, 40]. Moreover, there are several hints of different prescribing patterns of neurologists and GPs [25, 38]. Ruof et al. found that neurologists prescribed cholinesterase inhibitors to 44.6% of their Alzheimer patients, while GPs only treated 9.0% of their patients with cholinesterase inhibitors [41]. Thus, physician specialisation seems to be a reasonable indicator for the prescription of any anti-dementia drug. A prescription of neuroleptics could indicate the presence of behavioural and psychological symptoms of dementia (BPSD) and these symptoms are more common in the middle or severe stages of dementia.

Factors related to the prescription of Ginkgo biloba

The prescription of Ginkgo biloba was positively associated with patient age, whereas the chance for the prescription of Ginkgo biloba was smaller than 1 for women, patient treatment by a neurologist, and patient diagnosed with Alzheimer's disease. Co-morbidities and antidepressant or neuroleptic therapy did not reveal a significant impact, which might be due to the small sample size. Our results at first oppose the results of Nahin [42], who found that use of herbal and dietary supplements in older people in the United States was associated with female sex, a higher income, and a higher MMSE score in multivariate analysis. Although a subgroup analysis of factors related to the use of Ginkgo biloba had shown only race (white vs. nonwhite) as a significant factor, our results indicate that male patients are more likely to be prescribed Ginkgo biloba. At first this seems to counter studies which show that females are more likely to use CAM. However, it has to be noted that in dementia patients, the spouse might be the communication counterpart of the physician, which in the majority of cases of male dementia patients can be assumed to be the women. Thus, the predilection for CAM might be transferred to the partner.

Most of the physicians who asked for Ginkgo biloba, according to van den Bussche [25], assumed the potential effectiveness of Ginkgo biloba: 85% of the neurologists and 59% of the GPs attribute a moderate to high effectiveness. Thus it is not surprising that in our survey Ginkgo biloba was mainly prescribed by a neurologist. This is underpinned by the survey of Hillmer [43] on prescribing practices of Canadian family physicians with regards to Alzheimer's disease. The survey found that perception and self-reported knowledge are the most important factors that significantly predict lower prescribing rates. In the case of vascular dementia, there is no recommendation in the guidelines. Compared to other forms of dementia and taking the results of Hillmer into account, one therefore would assume a higher application of ginkgo preparations, which indeed was given in our data.

Neuroleptics

The prescribing frequency of antipsychotic drugs in geriatric patients is commonly overestimated particularly with respect to the administration of benzodiazepines, which are not recommended (DEGAM, DGN/DGPPN) [44] due to higher mortality and a higher tendency to fall. Thus, both conventional and atypical neuroleptics should only be administered for a short period of time (DEGAM, DGN/DGPPN) [18, 19]. According to Meyer et al., 28.4% of 2367 residents in 30 German nursing homes received neuroleptics and 23.6% were prescribed anxiolytic or hypnotic medication, mostly butyrophenone [45]. These rates do not differ from the 56.4% antipsychotic drugs being prescribed to 888 residents in Munich nursing homes [46]. Similar results were obtained from Mann et al. 2009 in Austria (antipsychotics 45.9%) [47], while the 31% reported in Italy was slightly lower [36]. Our results of only 14% thus are comparably low and might be explained by the CAM-orientation of the physicians. However, it is striking that benzodiazepines were mainly prescribed by GPs, while there was no significant difference between GPs and neurologists in the administration of neuroleptics.

Concomitant use of anticholinergics and cholinesterase inhibitors

In recent years an intensive and still ongoing debate on the concomitant use of anticholinergics and cholinesterase inhibitors has emerged. It is assumed that anticholinergics may attenuate the effects of cholinesterase inhibitors. Other studies suggest that anticholinergics are prescribed to palliate side effects of cholinesterase inhibitors [48]. Data from registers from Australia and the USA reveal a proportion of joint prescriptions between 20 and 47% [49–51]. Our sample is too low to draw concrete conclusions about the concomitant use of anticholinergics and cholinesterase inhibitors; however, with only one of 40 patients being prescribed this combined treatment, one may assume that CAM-oriented physicians might be more cautions in the concomitant use of anticholinergics and cholinesterase inhibitors.

Limitations

The present study has several important limitations which should be taken into account when interpreting the results. Firstly, additional data on the dementia diagnoses are lacking, such as a functional evaluation with the Barthel Index or an evaluation of the cognitive status with the Mini Mental State Examination (MMSE). The lack of objective tests for dementia massively increases the possibility of misclassification bias with respect to dementia type and severity. The proportion of Alzheimer's disease on all types of dementia normally should be between 50% and 70%. This is not reflected in our study which only found a proportion of about 10%. One explanation might be that GPs mainly (57%) classified the dementia as "unspecific". Secondly, although physician prescribing data were subjected to an internal review, coding inaccuracies cannot be ruled out entirely. Thirdly, data on subsequent medication use in patients who switched physicians were unavailable. Finally our study only presents data from a very small percentage of CAM physicians in Germany. The participation of 22 out of a possible 362 possible physicians does not support a broad generalisability of our findings nor does it allow carrying out detailed subgroup analyses. Thus, this analysis as such, should be viewed as a first insight into this field. In addition the present study also lacks a direct comparison group. Further research on this subject therefore would benefit from including a broader sample of CAM physicians as well as a stratified comparison group of conventional primary care physicians.