Article Summary
Pelton and others (2008) published an article on a pilot study of donepezil augmenting effect to antidepressant treatment in elderly depressed patients with cognitive impairment. They selected the randomized double-blind controlled study design. In the introduction section, the authors focused on several points of interest; first, depression by itself is an important cause of suffering and disability in the elderly. Second, when there is medical comorbidity, depression worsens the quality of life outcome; third, previous studies displayed the common association of depression and cognitive impairment. Fourth, there is a lack of research on long-term treatment implications for both conditions. They pointed to a previous study on the effect of sertraline yielded limited success to control depression and cognitive impairment. Therefore, the aim of the present study was to evaluate the augmenting effect of donepezil for enhancing cognition in elderly patients suffering from both conditions. They selected 23 patients who met the DSM-IV criteria for major depression; however, they relied on subjective cognitive complaints and impaired performance measures as inclusion criteria for cognitive impairment. They ran the treatment trial in three stages, first an 8 weeks trial on sertraline alone, second, a 12 week of randomised double-blind placebo controlled trial of adding donepezil. The third phase of 32 weeks followed, where patients were given donepezil alone with monthly monitoring. Results showed that donepezil is advantageous to elderly patients in the cognitive domain. They suggested that adding a choilne esterase inhibitor to antidepressant treatment might improve the cognitive outcome (Pelton et al, 2008).
Since the authors did not clarify the theoretical framework of their study design, choilne esterase’s mechanism of action, drug-drug interaction and contraindications, the researcher conducted a search to be able to comment on the article and fill in the gaps.
Search methodology
The researcher conducted an internet search of the following database:
- NLM Gateway, a service of the US National Institutes of Medicine.
- PubMed, a service of the US National Library of Medicine and the National Institutes of Medicine.
- Medscape database.
- CDC: Centers for Disease Control and Prevention.
- National Center for Health Statistics.
- The Cochrane Collaboration database.
- Biomed Central database.
- Google scholar web site.
Terms of search about the study design were types of clinical study design, randomised trials, double blind study design, randomised double blind study design, placebo controlled study design, and pilot studies (all terms were tried using principles, basics, and fundamentals as prefixes). For the medical-drug part of the study, the same databases were used with the following search terms: cognition enhancers for treatment of depressed elderly patients, treatment strategies for depression with cognitive impairment in elderly patients. In addition, the following search terms were tried; Efficacy of sertraline for the treatment of Depression and cognitive impairment in elderly patients, Donepezil mechanism of action, effects and side effects, and Sertraline Donepezil drug-drug interaction. Articles selected were those published between 2001 to March 2009.
Article Critique
The term evidence-based, represents a revitalisation state of nursing and medical practice. The standard of best evidence attainment needs randomization of patients either to alternative treatments or to compare known treatment strategies to newly developing ones. Apart from ethical issues, obtaining a balanced state between what is established, and what a new study suggests or the level of uncertainty in results links strongly to the quality of study design (Oricha and Yauri, 2003). Since the authors did not explain the theory behind the study design selected, which is a point of weakness, filling this gap mandates a discussion on the selected study design.
Study design theoretical framework
The need for randomised studies in medical research is to determine the effects of a health care interference (diagnostic or therapeutic) effect when they are not completely obvious from observational studies (Smith, 2003). Randomised controlled studies are studies where patients (participants) are grouped arbitrarily, that are at random to receive one of several interventions. Random distribution reflects that participants’ allocation to one of the study groups by chance alone, which is not determined or influenced investigators, the clinicians, or the study participants’ choice. The objective of these studies is to measure and compare different effects (outcomes) either present or absent after the participants receive the interventions. Since the outcomes are measured, these studies are quantitative; in addition, because these studies are used to compare intervention, they are comparative in nature (Altman, 2006).
It is assumed that random distribution of participants, the participants’ traits are likely to be similar (baseline, or group balancing), although not always achieved, yet randomization in useful to isolate and quantify the study variables (interventions). In addition, it is useful to reduce the effect of individuals’ factors that might influence the outcome (confounding factors) (Kassirer, 2004).
There are two methods for randomization; first, restricted randomization that aims at keeping the number of participants in all groups close if not similar. Second is, the stratified randomization, which shows more concern to the characteristics of participants (regarding age, weight, comorbidity…) keeping these characteristics as close as possible in all groups.
Irrespective of the method used for randomization, there are three principles to follow, first is, the investigators should clearly outline the rules that control participants’ allocation. Second is, the rules outlined should be strictly all through the study course (Smith, 2007). Finally, during results analysis, data mining, which is the inapt search for statistically significant relationships in a data set whether intentional or unintentional, should be avoided. Data mining occurs because of improperly chosen statistical procedures and is made possible by statistical software. Therefore, data analysis should be theory-driven, and depends more on outcomes measured through participants’ cooperation (Peduzzi et al, 2002).
Blinding a randomised control study
Based on the above discussion, the four limbs of a randomized controlled study are the participants, investigators and data collectors-analysts and those who assess the study outcome (assessors). The term open label (non-blinded) studies points to trials where everyone included know who received which intervention throughout the study course. Single-blinded studies refers to one of the four limbs is unaware of the intervention all through the course of the study, while in double-blinded studies, participants, investigators and assessors are unaware of the intervention throughout the study course. Triple-blinded studies include, in addition, the data analysts being unaware of the intervention used (Schulz and Grimes, 2002). Blinding of randomised studies reduces the discrepancy of outcome assessment; besides, conformity and retention of the study’s participants, further, it reduces the bias resulting from additional care or treatment (Devereaux et al, 2001).
The question of placebo
Placebo is a substance without medical benefit used to compare the value of active medication or intervention (like psychological treatment). Therefore, in medical trials, the main objective of placebo studies is to ensure the validity of the study intervention used. There are ethical questions about placebo on the background of compromising enough quality care for some patients, therefore there are principles of placebo agreed upon, first is monitoring every patient included in the study. The second is early skipping of a patient who suffers adverse effects linked to lack of quality care medication, the third is shortening the study duration to a minimum and finally each participant should sign an informed consent. Despite the ethical debate, randomised controlled double blinded placebo studies are the gold standard for clinical trials research (Regan et al, 2005).
Theoretical explanation of how choline estrase medications act
Human and animal model studies showed the central cholinergic system to be involved in memory and cognition, and that centrally acting acetyl choilne agonists and antagonists increase or reduce cognitive functions. Changes in the central cholinergic system may occur because of aging or neurodegenerative diseases (like Alzheimer’s disease) and may take the form of decreased choilne precursor uptake, decreased acetylcholine synthesis, release or changes in ligand receptor binding (Atri and Sherman, 2004).
Choline acetyltransferase enzyme mediates acetylcholine synthesis, after synthesis, it is stored in vesicles located near the synaptic junction to be released on neurone activation (depolarisation). After release, the choline esterase enzyme is responsible for hydrolysis acetylcholine. Thus mechanisms of enhancing acetyl choilne action to improve memory and cognition; first is through administering precursors to acetylcholine synthesis (proved ineffective) or second through administering cholinesterase inhibitors (Cinar et al, 2008). Donepezil is a choline esterase (ChE) inhibitor that proved efficiency in improving memory and cognitive functions in patients with neurodegenerative disorders. It has a selective inhibitory action on AChE, unlike rivastigmine which inhibits both AChE and BuChE, and characteristically its action is reversible (Ellis, 2005).
Weaknesses and strengths
Questions on the authors’ arguments
Although the authors built up a good well-organized discussion in the introduction section leading to their study’s objective, yet in the discussion section, they provided results of some previous studies to compare their results. They did not provide a critique to previous literature with Meta analysis or a review study that might lead to a stronger inference.
The question of sample size
There are no predetermined or simple formulas to develop a proper sample size for a quantitative clinical trial (like the one at hand); however, the larger sample sizes the more representative the sample is and accordingly the smaller statistical error. A substitute methodology when sample size is small is to go for a pilot study (Salganik, 2006), this may be the reason of deciding the title as pilot study. In this study, small sample size led to large variance with the placebo and donepezil treatment groups. Finally, although the study population was homogenous; yet, the inclusion criteria for depression were heterogeneous, besides the authors did not provide analysis to Peterson’s criteria of cognitive impairment.
The question of information
The authors reported no side effects of donepezil treated patients; however, reports of side effects are many including nausea, vomiting, diarrhoea, muscle cramps, insomnia, and fatigability. These side effects are known to take place more common with 10 mg doses than those used by the authors in phases B and C (Albers and Reist, 2004).
The duration of the study (52 weeks) and being quantitative and longitudinal are points of strength in this study.
Notes on statistical methods
The authors did not comment on validity of their study. There are two domains of quantitative research, first is the scope or domain of theory that is the acquired knowledge of how actions take place (muscle contracts in response to a stimulus). The second is observations. Based on two domains there are four constructs (build-ups). First is the cause build up (construct), when a researcher starts an experiment there exists a theory of what the cause of a certain observation is. Second is the effect build up (construct) that is for an experiment there is a theory of what the cause may produce. However, each of these constructs has to turn into a figure, a measure, a method based on the experiment’s observations. Third is the conclusion build up (construct), which is the answer to the question whether there is a correlation between the experiment variables or not. Internal validity relates to the conclusion construct that is the relation between variables a relationship of cause and effect or not. Fourth, construct validity supposes there is a cause effect correlation, does the treatment reflect the theory of using it; besides, did the response reflect the hypothesis of what is expected. External validity relates to constructing validity in the context of can the results of this particular experiment be generalized under different circumstances (Trochim, 2006).
They also did not comment on the descriptive statistics of their data. Descriptive statistics have three functions: 1- To measure the central tendency of data that is how the data collected aggregate around a most typical value. Three values measure the central tendency of data: a- the mean, which is the average value for a column of data, b- the median, which is the middle observation of a column of data (the largest value of the smaller half of data). Besides, c- The mode, this is the most frequent value in a column of data. The mean is the most significant because of its sensitivity (affected by changes of the sample size) (Robson, 2002). The second function of descriptive statistics is to measure data dispersion (data spread or variability around the mean). Three measurements are commonly used for this purpose: A- the range: It is the difference between the highest and lowest value in a column of data. B- Standard deviation, which quantifies data inconsistency around the mean. C- Variance, it is an expression of the observed differences from what would be expected if there were no natural variation (the same as the standard deviation). It is the mean of squared deviations from the mean and is used when sign (- or +) is considered (Robson, 2002).
Conclusions
Cognitive comorbidity to depression affect the quality of life of depressed patient as it affects their health outcome. The search for medications to improve cognitive functions in depressed patients is continuous, in this context; cholinesterase inhibitors like donepezil represent a promising adjuvant line of treatment. However, standardised clinical trials with larger sample sizes are still needed to reach a conclusion, which may be helped by Meta analysis studies of the previous literature.
References
Albers, LJ and Reist, C, 2004. Handbook of Psychiatric Drugs. Laguna Hills, CA: Current Clinical Strategies Publishing.
Altman, DG, 2006. Practical Statistics for Medical Research. 2nd. London: Chapman & Hall.
Atri, A and Sherman, S, 2004. Blockade of Central Cholinergic Receptors Impairs New Learning and Increases Proactive Interference in a Word-Paired Associate Memory Task. Behavioral Neuroscience, 118(1), 223-236.
Cinar, R, Freund, TF, Katona, I, et al, 2008. Reciprocal inhibition of G-protein signaling is induced by CB1 cannabinoid and GABAB receptor interactions in rat hippocampal membranes. Neurochemistry International, (52), 1402-1409.
Ellis, JM, 2005. Cholineesterase Inhibitors in the Treatment of Dementia. JAOA, 105(3), 145-158.
Devereauz, PJ, Manns, BJ, Ghali, W.A et al, 2001. Physician Interpretations and Textbook Definitions of Blinding Terminology in Randomized Controlled Trials. JAMA, (285(15)), 2000-2003.
Kassirer, JP, 2004. On the Take: How Medicine’s Complicity with Big Business Can Endanger Your Health. USA: Oxford University Press.
Oricha, BS and Yauri, MB, 2003. Uncertainty Principle versus Clinical Equipoise in Clinical Trials in Sub-Saharan Africa: Are they Really Tenable. Annals of African Medicine, 2(2), 99-100.
Pelton, GH, Harper, OL, Taber, MH, et al, 2008. Randomized double-blind placebo-controlled donepezil augmentation in antidepressant-treated elderly patients with depression and cognitive impairment: a pilot study. Int J Geriatr Psychiatry, (23), 670-676.
Peduzzi, P, Henderson, W., Hartigan and Lavori, P, 2002. Analysis of Randomized Controlled Trials. Epidemiological Reviews, (24(1)), 26-38.
Regan, J, Hamer, G, and Wright, A, 2005. The Placebo: Promise and Compromise. Tennessee Medicine, 138-139.
Robson, C, 2002. Real World Research: A Resource for Social Scientists and Practitioner-Researchers (Regional Surveys of the World). Chicago, Illinois: Blackwell Publishing Limited.
Schulz, KF and Grimes, DA, 2002. Blinding in randomised trials. The Lancet, (359), 696-700.
Salganik, MJ, 2006. Variance Estimation, Design Effects, and Sample Size Calculations for Respondent-Driven Sampling. Journal of Urban Health, 83(7), i98-i112
Smith, R, 2003. Medical journals and pharmaceutical companies: uneasy bedfellows. BMJ, (326), 1202-1205.
Smith, R, 2007. The trouble with medical journals. London: Royal Society of Medicine Press.
Trochim, WMK, 2006. Introduction to Validity.