The purported benefits of aromatherapy are so taken for granted in our society that aromatic products, whether truly infused with essential oils or not, abound on store shelves. Choose from an assortment of candles, bath and beauty products, or diffusers for home and office to cure what ails you. The claimed effects run the gamut from mood enhancement to specific healing properties. While acknowledging that many of these benefits have historical affirmation from a long lineage of use, when it comes to integrating alternative therapies with modern medicine, we are somewhat bound by the rigors of scientific inquiry to prove their worth.

The inclusion of essential oils in massage treatment is by and large a well-accepted practice and for some therapists, an integral part of their service. For the purposes of this column, we put aside the separate, but certainly critical, aspect of proper training as developed over centuries by healers of all origins. Instead, we focus on the ever-present dilemma of establishing within scientific construct the validity of complementary and alternative therapies (CAM). Not only that they do indeed provide healing, but also the conditions for which they are best suited. And so it is with aromatherapy.

With such an exuberant consumer population, defining clear-cut recommended use and benefits of specific essential oils seems to have its merits. But aromatherapy studies are few and far between, and some in the CAM field even question their value. In 2000, a systematic review of aromatherapy by Cooke and Ernst was presented in the British Journal of General Practice. Accompanying this paper was an editorial comment by Andrew Vickers of the Integrative Medicine Service at Memorial Sloan-Kettering Cancer Center in New York City, titled “Why aromatherapy works (even if it doesn’t) and why we need less research.” Both the review and Vickers’ comments bring up a host of questions regarding the validity and usefulness of aromatherapy research.

Research Reviews

Cooke and Ernst were looking for randomized controlled trials (RCT) that would substantiate the clinical benefits of aromatherapy. The 12 studies qualifying as RCTs were all small and flawed, but the team conceded the results do support a modest, short-term effect for anxiety reduction (anxiety being the endpoint measure of more than half of these RCTs). “However,” they note, “the data does not support a hypothesis that there may be legitimate clinical indications for the prescription of aromatherapy massage in a healthcare setting; it seems to have no lasting effects, good or bad.”¹ The concern here is cost-effectiveness for a therapy that has little scientifically-proven staying power or relevance to specific conditions.

Several problems arise in addressing aromatherapy research, the first being that aromatherapy is frequently administered via massage, which has already been proven to reduce anxiety. As Vickers notes, perhaps studying this combination is an inappropriate application of research time and funding.² And with Cooke and Ernst’s assessment of only a modest effect when aromatherapy is added, this may be a point worth considering.

A second conundrum of fitting aromatherapy into scientific standards is compounded by inexperienced researchers. The small qualifying literature base identified by Cooke and Ernst had many weaknesses throughout, including lack of an expert statistician to evaluate data, no clear hypothesis, and small sample size, which decreases statistical power. These concerns are not outside the norm in the current state of CAM research. But other factors certainly demand specific attention in future work.

In those studies reviewed, there was no referenced justification for the type of essential oil chosen. The authors also question the feasibility of double-blinding a study in which the odor of aroma cannot be masked. “Thus the question of whether we are dealing with specific or non-specific effects may never be resolved completely satisfactorily,” they write. The waters are muddied even further by a potential psychological response to odors associated with pleasant memories.

The team also points to the difficulty of determining mechanism of action as “there was no attempt to differentiate between the effects of any transdermal absorption of the oils and the effects of smell.” Overall, the team reports these studies are so significantly flawed as to “prevent firm conclusions from being drawn. The effects under investigation are not easy to measure, and their size seems likely to be small.”³

In past columns here, we have noted cautionary advice by experts that research is best conducted by those who have no financial investment in the final results. (And, of course, it is duly noted that much pharmaceutical research is indeed sponsored by product manufacturers.) Vickers points out that in each trial, the first author “is a practicing aromatherapist, often a high profile advocate (e.g., writes aromatherapy books, teaches nurses) and generally not an experienced researcher with a significant number of publications.”⁴ In such cases, not only is there the possibility of bias affecting outcome, but also lack of qualified researching experience to ensure methodological accuracy. Even more recently, a systematic review by Thorgrimsen et al. (2003) examined the use of aromatherapy as an intervention for dementia (related to cognitive function, quality of life, and relaxation). Unfortunately, they found only two studies qualifying as RCTs and of these only one (Ballard et al. 2002) had usable data for analysis.⁵ In the Ballard study, application of lemon balm (Melissa officinalis) was compared with use of placebo oil to decrease agitation in patients with severe dementia. The 72 subjects randomly assigned to experimental and control groups of equal number were given twice-daily care over a four-week period. Of 71 subjects completing the trial, reduction in agitation was shown in 60 percent of the lemon balm group and 14 percent of the placebo group. Additionally, aromatherapy treatment subjects fared better in increased quality of life scores.⁶

Although further analysis from Thorgrimsen’s team confirmed the significance of these results, they point out the presence of several methodological flaws. “More well designed large-scale RCTs are needed before conclusions can be drawn on the effectiveness of aromatherapy,” they write, and researchers should investigate “whether different aromatherapy interventions are comparable and the possibility that outcomes may vary for different types of dementia.”⁷

Imperfections aside, it is frequently just this type of work that provides inspiration for the real breakthroughs. Dementia, a growing concern within our elderly population, has devastating impact on both patients and their loved ones. Although, as Vickers suggests, we need to be attentive in prioritizing medical research and funding, for these families no application is too small to be of service.

Lessons to Learn

Learning the hard way!” begins the title of a 2003 report by Westcombe et al., British researchers who describe their experiences conducting an aromatherapy massage RCT for cancer patients. Their challenges mirror concerns already highlighted, and then some. Referring to the Cooke and Ernst review, Westcombe et al. introduce their work by stating, “The present study aimed to address some of these criticisms and take up the challenge of providing a scientifically rigorous evaluation of a complementary therapy in a study population compromised in terms of their physical health.”⁸

The obstacles encountered by the team were significant enough to call for design modifications as the trial progressed, although authors note these changes “were consistent with the original basic study aims and design principles.”⁹ Small sample size (and subsequently underpowered results) has been a recurrent problem in previous aromatherapy research. The Westcombe team, valiantly seeking to establish a study group of 508 participants, soon encountered recruitment difficulties. Criteria established for the study originally included only those patients with advanced cancer who were also experiencing “mild to moderate levels of psychological distress.” To improve recruitment, the trial was opened to all cancer patients, irrespective of disease stage.¹⁰

A second modification involved the removal of a relaxation therapy comparison group, although the primary control (no intervention) group remained in place. The original goal of 508 subjects was amended to 258 and another recruitment center was added to the four already being utilized. At the time of the authors’ writing, recruitment had reached an adequate number where data collection could be completed successfully.¹¹

In reflecting on their journey, Westcombe’s team outlines the “lessons learned.” Included among points to consider in maintaining realistic design expectations, they encourage flexibility in setting exclusion/inclusion criteria.¹² (While this can indeed enhance recruitment, some CAM studies are criticized for being too heterogeneous in their subject base and thus lose ground in applicability.) The team also recommends implementing short, follow-up periods and flexible methods of data collection, such as telephone contact and home visits, to avoid participant attrition.¹³

Preliminary investigation prior to embarking on the full RCT is crucial, the team writes, to “highlight potential obstacles to success.”¹⁴ Qualitative and observational studies can not only flush out glitches but also build a bridge of trust and acceptance within prospective study settings. Pilot projects are commonly used in research to address feasibility and relevance for larger projects.

From inception to completion of data collection, maintaining a high profile is also important. The authors suggest a proactive approach — encouraging what they call “clinical champions” at each data collection site — to stimulate referral and reinforce the relevance of the work. To keep on-site therapists from being isolated and to ensure their adherence to treatment protocol, authors recommend team meetings to discuss study-related issues. Likewise, data collectors can benefit from supervision aimed at decreasing their isolation and assisting them in managing patients who may be of concern.¹⁵

Although recruitment problems necessitated changes in study design and prolonged the trial period beyond that originally funded, the Westcombe team describes the final result as “a pragmatic randomized controlled trial with a comprehensive measurement strategy.”¹⁶ Time will tell if their efforts are not only reported in some future systematic review, but also contribute positively to the evaluation of aromatherapy massage.