EXTENT OF CM USE

According to a number of hospital surveys conducted in the United States and in Australia, use of complementary medicines (CMs) is well established among surgical patients (Braun et al 2006a, Grauer 2004, Kaye et al 2000, Leung et al 2001, Norred 2002, Norred et al 2000, Silverstein & Spiegel 2001, Tsen et al 2000).

One survey of 3106 patients attending preoperative clinics in the USA found that 51% were using CMs, and that women were more likely users than men (Tsen et al 2000). Age was also a factor, with patients aged between 40 and 60 years greater consumers of CMs. Overall, only 21% of patients sought advice from a healthcare provider about supplement use, whereas the majority cited friends and family or their own decision as reasons for taking these medicines.

Another survey of 1017 patients presenting for pre-anaesthetic evaluation found that approximately one in three were taking one or more herb-related compounds (Kaye et al 2000). Disturbingly, nearly 70% of these patients did not report use when they were asked during routine anaesthetic assessment. Lack of communication was also detected in a review of 2560 patients, which identified that 39% were taking natural supplements, with herbal medicines being the most popular (Leung et al 2001). Of these, 44% did not consult with their primary-care physician and 56% did not inform their anaesthetist before surgery.

Similar trends have also been identified among Australian surgical patients (Braun et al 2006a, Grauer 2004). A survey of 508 presurgical inpatients at two tertiary hospitals found that in the 2 weeks before admission, 46% had used CMs, of which 38% were self-prescribed (Braun et al 2006a). Just over half these patients planned to continue using them while in hospital. During the 2 weeks before admission, 64% had not discussed use with any community-based or hospital-based healthcare professional. Similar to findings of other studies, patients using CMs were most likely to be female, and to have pursued higher education and earn above-average income. Interestingly, this study found that age was not predictive of use, and that older patients were just as likely as young patients to be using CM. Importantly, 58% of surgical patients did not discuss their use with a hospital doctor, nurse or pharmacist. When patients were asked why they did not tell hospital staff, it was not because they feared being judged negatively, but overwhelmingly because they were not asked. This was also the first Australian study to ask patients about their interest in hospital-based CM therapies; 85% expressed an interest.

Previously, a study that evaluated preoperative use of CM products at two other hospitals had found that 20.4% had used vitamin supplements and 14.3% herbal medicines in the previous 6 weeks, but only 28% had informed hospital staff about use (Grauer 2004). Once again, the majority of patients had self-selected their treatment.

These findings pose several important questions, such as whether surgeons, anaesthetists, physicians or pharmacists are aware of the potential safety issues associated with CMs, routinely ask patients about herbal and natural medicines in a way that encourages open dialogue, and have the knowledge and resources to interpret the answers. Survey results published in 2001 found that, although physicians did ask about herbal and natural medicine use, most did not check with reference texts for more information (Silverstein & Spiegel 2001). Disturbingly, the survey also detected an obvious lack of knowledge about common herbal medicines among physicians. A more extensive study was conducted recently with hospital doctors and pharmacists at an Australian tertiary hospital (Braun et al 2006b). The study of 127 surgeons, anaesthetists, physicians and pharmacists found that 68% thought CMs could be dangerous and that patients’ use needs to be monitored; however, only 28% routinely asked patients about CM use. All groups lacked knowledge and confidence in dealing with CAM, and 81% felt they had insufficient knowledge to be able to identify whether a CM product could adversely affect patient care. Despite this, few checked for side effects or interactions when patients using CMs were identified.

When patients do report using herbal and natural medicines, hospital doctors and pharmacists do not routinely document this information in patients’ charts, making effective communication even more difficult (Braun et al 2006b, Cohen et al 2002 ). Cohen et al (2002) found the prevalence of supplement use was 64% in a group of 212 patients, but only 35% of all supplements were recorded in patients’ charts.

ASSESSING POTENTIAL FOR ADVERSE EFFECTS

The use of CAM by surgical patients should not pose any serious concern, unless they have the potential to cause adverse reactions that would negatively affect the outcome of surgery and impede recovery. Potential unwanted effects such as increased bleeding risk, sympathomimetic effects or detrimental interactions with medicines commonly used in the perioperative period are some examples.

Two studies conducted in the United States have investigated this further and suggest that some CM products used by surgical patients have the potential to alter bleeding risk, cardiovascular stability and interact with anaesthetics. A survey of 500 elective surgical patients found that 51% used CAM products during the 2 weeks before surgery, and that 27% of CMs taken had the potential to inhibit coagulation, 12% to affect blood pressure, 9% to cause sedation, 5% to have cardiac effects and 4% to alter electrolytes (Norred et al 2000). A larger survey of elective outpatients from 16 hospitals in the United States found that 67% used herbal and natural supplements or homoeopathic medicines and, of these, 34% used medicines that had the potential to interact with anaesthetics or to inhibit coagulation (Norred 2002). A review of the medicines used identified garlic (9.5%), cranberry (5.2%), ginkgo (4.6%), ginseng (4.2%), echinacea (4.1%) and products containing ephedra (3.2%) as the most commonly consumed herbal medicines, whereas multivitamins (37.5%), vitamin E (23.3%), calcium (21.5%) and vitamin C (21.4%) were the most commonly taken nutritional supplements.

Although useful, the results obtained by these two papers should be interpreted carefully because much of the safety information presented is speculative and based on studies conducted in other populations or test models. Close inspection of the lists of herbal and natural supplements claimed to inhibit coagulation reveals that no distinction is made between those medicines with clinically observed effects and those with predicted effects based on in vitro or in vivo evidence. Furthermore, the majority of claims made about adverse reaction potential are based on case reports and in vitro and in vivo tests, and not on controlled clinical trials. Without documenting where evidence comes from for each claim, readers can easily give equal weight to all claims, making the lists of limited clinical value and likely to overstate the risk. Additionally, both papers fail to state full botanical names when referring to herbal medicines and are open to misinterpretation.

Developing a list of herbal and natural supplements with a realistic potential to cause adverse effects in surgery is difficult at the current time, as the pharmacology of many of these medicines, including pharmacodynamic and pharmacokinetic properties, has yet to be fully elucidated. Until evidence from controlled clinical studies is available, other evidence will be used but should be interpreted cautiously.

SAFETY, SIDE EFFECTS AND INTERACTIONS WITH OTHER MEDICINES

The risk of harm associated with the use of CMs is largely unknown; however, it is likely to be greatest in high-risk surgery; that is, when the surgical procedure to be undertaken and the patient’s health status put the patient at real risk of serious complications. Factors such as preexisting fluid and electrolyte status, cardiorespiratory performance, comorbidities and medical pretreatments need to be considered when assessing safety in patients (Girbes 2000). Additionally, the type of surgical procedure should be considered, in particular those procedures in which increased bleeding would be a serious complication (e.g. orthopaedics or neurosurgery) or put significant physical strain on the patient (e.g. coronary bypass) would be considered high risk.

BLEEDING RISK

Surgery is often associated with some degree of blood loss, but this is usually limited by the body’s haemostatic mechanisms, which finely regulate interactions between components on the blood vessel walls, circulating platelets and plasma proteins. A retrospective survey of adverse surgical outcomes in several US hospitals found that postoperative bleeding accounted for 10.8% of all surgical adverse events, making it the third most frequent adverse surgical event (Gawande et al 1999). In practice, whether bleeding risk is a serious concern is usually dictated by the type of surgery to be undertaken. For instance, minor dental procedures would not be as seriously affected by increased bleeding as neurosurgery.

A number of explanations may account for increased bleeding, such as undiagnosed clotting disorders, liver disease or the preoperative use of certain medicines such as antiplatelet agents. Over time, it has become recognised that some CMs also have the potential to influence the body’s haemostatic response. Initially evidence was derived from case reports suggesting CM use was associated with bleeding or from in vitro tests indicating antiplatelet activity; however an increasing number of clinical studies are now being published to more rigorously determine which CAM products cause clinically significant bleeding.

Herbal medicines and food supplements

Whether preoperative use of herbal medicine alters bleeding risk is difficult to determine because few controlled trials are available for this specific population. Evidence is usually extrapolated from other sources to identify those medicines with suspected antiplatelet or anticoagulant activity and, while still largely speculative, this provides a guide until more robust evidence is available.

Difficulties arise when evidence is contradictory, making clinical recommendations even more difficult. The herb Tanacetum parthenium, also known as feverfew, provides a good example. Several test-tube studies and animal models have observed inhibition of platelet aggregation (Heptinstall et al 1988, Jain & Kulkarni 1999, Makheja & Bailey 1982); however, no significant effects were seen in a small study of 10 patients receiving feverfew (Biggs et al 1982). The problem of inconsistent information can add to practitioners’ confusion, as some databases are infrequently updated or give undue weight to case reports and seemingly less consideration of safety information derived from clinical studies. The case of ginkgo biloba illustrates this point.

Ginkgo biloba

Ginkgo biloba is a popular herbal medicine used mainly for cognitive decline and peripheral vascular diseases, supported by good evidence. Concern over whether G. biloba significantly increases bleeding first arose in response to several case reports describing haemorrhage during or after surgery (Hauser et al 2002, Schneider et al 2002) and evidence that one of its components, ginkgolide B, is a platelet-activating factor antagonist (Smith et al 1996). In the following years, at least 10 clinical studies have been conducted that have found no evidence of significant bleeding or platelet effects due to G. biloba ingestion (Bal Dit et al 2003, Engelsen et al 2003, Kohler et al 2004, Jiang et al 2005, Wolf 2006, Beckert et al 2007, Gardner et al 2007, Aruna & Naidu 2007, Lovera et al 2007, Carlson et al 2007). Studies have included young healthy volunteers, older adults, people with multiple sclerosis and people using warfarin or aspirin at the same time as G. biloba. An escalating-dose study found that 120 mg, 240 mg or 480 mg given daily for 14 days did not alter platelet function or coagulation (Bal Dit et al 2003). Only one clinical study demonstrated that EGb761 (80 mg/day) produced a significant reduction in blood viscosity after 30 days’ treatment (Galduroz et al 2007). When the available body of evidence is considered, it appears unlikely that G. biloba causes a significant bleeding risk to most patients. Unfortunately, a number of popular electronic databases and reference texts still cite bleeding risk as a major concern with its use, particularly when used in combination with warfarin (Braun 2008).

Food supplements contain concentrated forms of dietary foods, such as fish oils or fibre. Herbal and food supplements overlap in many cases because various herbs are also eaten as foods (e.g. ginger). Keeping this in mind, it is not unusual to find evidence suggesting that normal dietary intake of a food does not appreciably alter bleeding risk, whereas ingesting concentrated products will. For example, 4 g ginger daily does not alter platelet aggregation or fibrinogen levels according to one clinical study, whereas a dose of 10 g/day significantly reduces platelet aggregation according to another (Bordia et al 1997).

Commonly used herbal medicines and food supplements available OTC that have been found to inhibit platelet aggregation under clinical test conditions include garlic, ginger root, onion, policosanol and pine bark extract (Araghi-Niknam et al 2000, Arruzazabala et al 2002, Bordia et al 1975, 1997, Harris et al 1990, Jung et al 1991). Often, very high doses above normal dietary intakes are required to produce these results.

Fish oils

Fish oils supplements are extremely popular and are considered by some healthcare professionals to increase the risk of bleeding in surgery. A search through the Medline database reveals several case reports where bleeding episodes are attributed to fish oil ingestion (Buckley et al 2004, Jalili & Dehpour 2007, McClaskey & Michalets 2007). In each case, the person affected was elderly and also taking warfarin. One was a report of an elderly man taking high-dose omega-3 fatty acids (6 g/day) with both aspirin and warfarin who developing a subdural haematoma after a minor fall (McClaskey & Michalets 2007). Another case is reported of a 67-year-old woman taking warfarin for 1.5 years who doubled her fish oil dose from 1000 to 2000 mg/day, causing an associated elevation in INR from 2.8 to 4.3 within 1 month (Buckley et al 2004). A third case was of a 65-year-old male who had been taking warfarin for 6 months and then was recommended trazodone and fish oils, causing his INR to rise to 8.06 (Jalili & Dehpour 2007).

Although these case reports suggest omega-3 fatty acids interact with warfarin and increase the risk of bleeding, several intervention studies have come to a different conclusion. One randomised study of 511 patients taking either aspirin (300 mg/day) or warfarin (INR 2.5–4.2) found that a dose of 4 g/day of fish oils did not increase the number of bleeding episodes, bleeding time or any parameters of coagulation and fibrinolysis (Eritsland et al 1995). A smaller placebo-controlled study by Bender et al (1998) of patients receiving chronic warfarin therapy found that fish oil doses of 3–6 g/day produced no statistically significant difference in INR between the placebo lead-in and treatment period within each group. There was also no difference in INR between groups.

Most recently, Harris (2007) examined 19 clinical studies that used doses of fish oil varying from 1 g/day to 21 g/day in patients undergoing major vascular surgery (coronary artery bypass grafting, endarterectomy) or femoral artery puncture for either diagnostic cardiac catheterisation or percutaneous transluminal coronary angioplasty. Of note, in 16 studies patients were taking aspirin and in 3 studies patients were taking heparin. The review concluded that the risk of bleeding was virtually non-existent. Frequent comments accompanying the studies were ‘no difference in clinically significant bleeding noted’ or ‘no patient suffered from bleeding complications’. The same conclusion was reached in a 2008 review that stated no published studies have reported clinically significant bleeding episodes amongst patients treated with antiplatelet drugs and fish oils (3–7 g/day) (Harris et al 2008). In light of this body of evidence, it is apparent that fish oil supplementation does not significantly increase bleeding risk.

Emerging evidence further suggests that administration of fish oil to presurgical patients before cardiothoracic surgery has significant benefits, such as reduced rate of postoperative atrial fibrillation (Calo et al 2005). Such benefits have major implications for patients, hospital administrators and hospital waiting lists.