The use of traditional Chinese herbal medicines (TCHMs) among the presurgical population is widespread, but their impact on perioperative patient care is unclear. The authors estimated the incidence and risk of TCHM-related perioperative events.
In a Hong Kong cohort study, 601 patients undergoing major elective surgery were asked about their Western medicine and TCHM use in the 2 weeks before surgery. Unanticipated perioperative events were noted by attending anesthesiologists, blinded to patients' use of specific TCHMs. Modified Poisson regression models were used to obtain the relative risk of combined endpoints of perioperative events associated with TCHM use.
Of the 601 patients, 483 patients (80%) took self-prescribed TCHM, and 47 (8%) took TCHM by prescription (with or without self-prescribed TCHM) in the 2 weeks before surgery. The crude incidences of any combined endpoints of preoperative, intraoperative, and postoperative events were 23% (95% confidence interval, 19-26%), 74% (95% confidence interval, 71-78%), and 63% (95% confidence interval, 59-66%), respectively. Compared with nonusers, patients who took TCHM by prescription were more likely to have a preoperative event (adjusted relative risk, 2.21; 95% confidence interval, 1.14-4.29). The authors present four case reports to highlight the effect of TCHM by prescription on prolonged activated partial thromboplastin time and hypokalemia in the preoperative period. In contrast, there was no significant association between the use of any type of TCHM and the occurrence of either intraoperative or postoperative events.
The use of TCHM by prescription near the time of surgery should be discouraged because of the increased risk of adverse events in the preoperative period.
TRADITIONAL Chinese herbal medicines (TCHMs) are increasingly used throughout the world because they are considered to be effective and to have few side effects.1Despite well-established Western-style healthcare systems, the use of TCHM is highly prevalent in Asian countries2,3and among presurgical patients.4TCHM emphasizes a holistic approach to care, with recognition that diet is important in promoting health, as well as treating or preventing diseases.5Compared with Western medicine, most patients believe that the use of TCHM is more effective for health promotion (82%) than for treating serious illness (10%).4
Recent reviews6,7have highlighted the potential harmful effects of some herbal medicines in the perioperative period. The most common adverse effects of herbal medicine relevant to anesthesiologists and surgeons include impaired coagulation, cardiovascular side effects, electrolyte disturbances, and prolongation of the effects of anesthetic agents.5Because patients do not often disclose their herbal use, both anesthesiologists and surgeons are often unable to make specific associations between potential drug and herb interactions. Based on limited pharmacokinetic data, the current recommendation is that all herbal medicines should be ceased 2 weeks before surgery.5,7
Little information is available about the epidemiology of TCHM-related perioperative events in clinical practice. No studies to date have examined the clinical outcomes of surgical patients taking TCHM within 2 weeks of surgery. As a general principle of using TCHM, herbal remedies used for health promotion and minor ailments tend to be mild in nature and relatively harmless.4TCHMs prescribed by registered traditional Chinese medicine practitioners tend to be more pharmacologically active and thus are more likely to cause adverse reactions than nonprescription TCHMs.4When the same herbal ingredients are prepared as a decoction, as usually in the case of a TCHM by prescription, it is more potent and produces a faster onset of action than if it were administered as a tablet.8Therefore, we conducted a cohort study to determine the incidence and risk of TCHM-related perioperative events. We hypothesized that the use of TCHM prescribed by registered traditional Chinese medicine practitioners would be associated with an increased risk of a perioperative event. We expected that this research would assist in the development of guidelines and interventions designed to reduce the risk of TCHM-related events in patients undergoing surgery.
Materials and Methods
Study Setting and Population
The study was conducted at a teaching hospital affiliated with The Chinese University of Hong Kong from January 2003 to September 2004. Patient recruitment was suspended during the severe acute respiratory syndrome outbreak in Hong Kong (March to June 2003). The project was approved by the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee and the Survey and Behavioral Research Ethics Committee, Shatin, Hong Kong, China. All patients gave written informed consent. This study included all persons undergoing major elective surgery as defined in the Hong Kong Government Gazette .9All types of major surgery were included, except liver transplantation, emergency surgery, and cesarean delivery. Patients who had undergone a previous surgical procedure within 14 days or had a history of cognitive impairment were excluded from the study.
Patients were selected from designated operating theaters each week according to a computer-generated random table. They were followed up prospectively from the time of hospital admission through to discharge. A total of 633 patients were approached for consent. Thirty-two patients were excluded from the study (extent of the surgery was misclassified in 12, wrong recruitment from designated operating theater in 5, previous surgery within 14 days in 5, refused consent in 4, surgery cancelled before data collection in 2, local anesthesia without requiring the presence of an anesthesiologist in 2, and cognitive impairment in 2). Of the 601 remaining, intraoperative and postoperative data were not collected in 9 patients because their surgery was cancelled due to administrative reasons. A total of 592 patients participated in and completed the study.
Exposure to Herbal Medicines
All surgical patients were admitted as inpatients on the day before surgery. Patients were interviewed by one of the investigators (A.S.C.L.) on the day before their surgery using a standardized questionnaire. The first section of the questionnaire dealt with patients' general characteristics (age, sex, level of education, monthly income) and medical history. The patient's baseline blood pressure and heart rate were calculated as the mean recordings taken during the first day of hospital stay. In the second section, we asked patients about their regular use of Western medicine and TCHM within the 2 weeks before admission. We noted the form (herbal teas, herbal soups, self-prescribed prepacked over-the-counter preparations, and TCHM by prescription) and frequency of TCHM intake. TCHMs prescribed by a registered traditional Chinese medicine practitioner included a mixture of dried fresh herbs that had to be boiled in water, a powder that had to be mixed in boiled water, tablets, or a topical paste. Patients were provided with a list of 45 commonly used Chinese herbs, similar to a previous survey that examined TCHM use in the past 12 months.4In the third section, we asked whether the patients had seen a traditional Chinese medicine practitioner within the 2 weeks before admission and about any herbal prescriptions they had received. The time period was chosen according to current recommendations about when to stop herbal medicines.5,7Patients were encouraged to provide a copy of the TCHM prescription after the interview. The attending anesthesiologist was blinded to the patients' detailed TCHM history that was obtained from the interview. We classified level of TCHM use into three groups: (1) nonusers, defined as patients with no TCHM use; (2) self-prescribed users, defined as patients taking herbal teas, herbal soups, and over-the-counter TCHM preparations but no TCHM by prescription; and (3) TCHM-by-prescription users, defined as patients taking herbal prescription prescribed by a traditional Chinese medicine practitioner. Some patients in this group took additional herbal teas, herbal soups, and over-the-counter TCHM preparations.
The primary outcomes of the study were unanticipated perioperative events occurring within the 14 days before surgery to 24 h after surgery. Perioperative events were categorized as preoperative, intraoperative, and postoperative. For all patients, the attending anesthesiologist was asked to record the unanticipated occurrence of any of the perioperative events listed in table 1on a standardized data collection form. Abnormal laboratory results were defined as any result outside the reference range. Duration of stay in the hospital and hospital mortality data were also recorded. Hemodynamic data were downloaded to a computer from the physiologic monitors at 5-s intervals using software developed within our department. Data from the patient's anesthetic record were cross-checked against the data collection form for all patients, irrespective of whether a perioperative event had occurred.
For case reports, several reference books10–12and original journal articles were searched to obtain evidence of known pharmacologic effects of individual TCHMs. The diagnosis of “probable” TCHM-related events was made independently by two anesthesiologist reviewers (P.T.C. and C.S.T.A.) using the Naranjo tool.13The Naranjo tool is a 10-item scale13that considers previous reports of known drug effects, temporal relation between drug exposure and event, dose response, rechallenge response, objective evidence of the event, and alternative causes that might contribute to the event occurring. A score between 5 and 8 is regarded as “probable” causation. The maximum score possible in our study was 7 because it was not possible to rechallenge the patient to go through the whole process. This tool has been used in several case reports of adverse events related to herbal medicines in the medical literature.14–16This approach for determining the causal relation between herbs and adverse events is similar to that used by an advisory team on herbal safety from the Drug and Poisons Information Bureau and the Hospital Authority Toxicology Reference Laboratory in Hong Kong17and by the World Health Organization.∥
Continuous variables were expressed as mean (± SD) or median. One-way analysis of variance was used to compare the means across different patient groups. Categorical variables were expressed as proportions and compared using a chi-square test. The comparison of groups for hospital duration of stay was performed using appropriate nonparametric tests. Ninety-five percent confidence intervals (95% CIs) were calculated for the incidence of events occurring in the preoperative, intraoperative, and postoperative periods. Modified Poisson regression models18were used to obtain the relative risk of combined perioperative events associated with TCHM use with adjustments to covariates (age, sex, American Society of Anesthesiologists physical status, type of surgery, and visit to a TCHM practitioner within the past 2 weeks). To check the adequacy of the modified Poisson regression model, the deviance residuals were estimated. If the model is correct, the majority (95%) of these residuals would be expected to fall between ± 2.19
A sample size of 545 was expected to produce a 95% confidence interval equal to the prevalence of TCHM use ± 3% when the estimated prevalence of TCHM use was 85%. Because 10% of patients would give poor or incomplete history of their herbal medicine use, we needed to recruit 600 patients to take this factor into account (nQuery Advisor 4.0; Janet D Elashoff 2000, Los Angeles, CA). The level of significance was set at P < 0.05. Analyses were performed using STATA 8.2 (Stata Corporation, College Station, TX).
Of the 601 patients, 483 (80%) took self-prescribed TCHM, 47 (8%) took TCHM by prescription (with or without self-prescribed TCHM), and 71 (12%) did not take any form of TCHM in the 2 weeks before surgery. The TCHM-by-prescription users were older (median, 50 yr; interquartile range, 42–60 yr) than self-prescribed users (median, 43 yr; interquartile range, 21–54 yr) and nonusers (median, 43 yr; interquartile range, 12–54 yr) (P < 0.01). The proportion of American Society of Anesthesiologists physical status II–IV was higher in the TCHM-by-prescription user group compared with the self-prescribed user group (63% vs. 44%, respectively; P < 0.01). The duration of anesthesia was longer in the TCHM-by-prescription users (median, 160 min; interquartile range, 105–240 min) than in nonusers (median, 120 min; interquartile range, 78–168 min) (P < 0.01). More patients in the TCHM-by-prescription user group (30%) took regular antihypertensive drugs (β blockers, calcium antagonists, angiotensin-converting enzyme inhibitors) than those in the self-prescribed user group (15%) (P < 0.01).
The most common forms of self-medicated TCHM taken by 525 patients were herbal soups (79%), herbal teas (37%), and prepacked over-the-counter preparations (33%). Of the 32 patients (5%) who took Western medications and TCHM by prescription concurrently within 2 weeks of surgery, one patient was taking metoprolol with TCHM with known antihypertensive effects (pilose Asiabell root, Atractylodes, American ginseng, hare's ear root).
Thirty-five patients (6%) had visited a traditional Chinese medicine practitioner within 2 weeks of surgery. Of the 47 patients (8%) who took a TCHM by prescription, only 27 knew of all or some of the herbal ingredients used in the prescription. The most common form of TCHM by prescription was by a mixture of dried fresh herbs that had to be boiled in water to make a decoction (34 of 47 prescriptions). Eleven patients provided a copy of the TCHM prescription listing the TCHM ingredients after the interview. The self-prescribed TCHMs were used for health promotion or for treating minor illnesses, whereas prescribed TCHMs were taken for treating major illnesses (table 2).
Incidence of Perioperative Events
There were no cardiac arrests. All patients were alive at hospital discharge. One patient had neck hematoma after undergoing carotid endarterectomy and required further surgery within 24 h to control bleeding and intensive care treatment. The crude incidence of specific types of perioperative events are given in table 3. No epidural or spinal hematoma was reported in the 133 patients taking some form of TCHM who underwent regional anesthesia. Overall, the crude incidence of combined preoperative, intraoperative, and postoperative events were 23% (95% CI, 19–26%), 74% (95% CI, 71–78%), and 63% (95% CI, 59–66%), respectively. There was no difference in the crude incidences of combined preoperative, intraoperative, and postoperative events among the TCHM patient groups (fig. 1). On multivariate analysis, patients who took TCHM by prescription were more likely to have a preoperative event (adjusted relative risk, 2.21; 95% CI, 1.14–4.29) than nonusers after adjusting for significant patient characteristics (table 4). There was no significant association between the use of TCHM and the occurrence of either intraoperative or postoperative events after adjustment for covariates (table 4). There was adequate fit for all models. Patients who had any perioperative events stayed in hospital significantly longer than patients who did not (median, 6.0 vs. 3.0 days; P < 0.01).
Preoperative international normalized ratio (INR) and activated partial thromboplastin time (aPTT) results were available in 54% (n = 323) of the study population because these tests were ordered by surgeons according to their practice. There was no significant difference in the incidence of prolong INR or aPTT among 45 patients taking ginger compared with 278 who did not (20% vs. 17% and 7% vs. 11%, respectively; P = 0.67 and P = 0.60, respectively). The incidence of prolonged INR or aPTT among 48 patients taking any type of ginseng (American, Asian, Korean) was similar to 275 patients not taking ginseng (17% vs. 17% and 8% vs. 11%, respectively; P = 0.89 and P = 0.64, respectively).
“Probable” Preoperative Events Caused by TCHM by Prescription
The characteristics of the four patients with “probable” preoperative events caused by TCHM by prescription are shown in table 5. Preoperative TCHM-related hypokalemia occurred in three patients, with one requiring oral potassium chloride supplements before surgery. There were no obvious reasons (such as the use of diuretics and presence of renal and gastrointestinal disorders) that might explain these patients' hypokalemia. Cinnamon twig, Loranthus, Atractylodes, and Achyranthes root were common herbs with diuretic effects used in TCHM prescriptions in patients with hypokalemia. TCHM-related coagulopathy occurred in one patient, requiring a change in anesthetic plan, additional preoperative hemostatic tests, and abandonment of epidural analgesia (table 5).
This study is the first description of the impact of patients taking TCHM near the time of surgery on perioperative care. In a population with a high prevalence of TCHM use, the crude incidences of events before, during, and after surgery were similar among nonusers, self-prescribed TCHM users, and TCHM-by-prescription users. However, after adjusting for differences between the three TCHM groups, we found a significant effect of TCHM by prescription on the risk of preoperative events. In particular, TCHM-by-prescription users were more than two times likely to experience hypokalemia or impaired hemostasis (as measured by prolong INR and aPTT) than nonusers in the preoperative period. In the same analysis, there was no difference in the risk of preoperative events between self-prescribed TCHM users and nonusers. The clinical significance includes that (1) additional interventions are often required to correct or monitor the TCHM by prescription–related adverse effects, and (2) it may necessitate a change in the anesthetic treatment of the patient as highlighted by the one of the case reports. However, despite the high prevalence of TCHM use in this surgical population, there was no clinically significant negative impact on intraoperative and postoperative care.
A major difference between the Western and oriental use of herbs is that Western herbs are more often used in an isolated form, whereas in traditional Chinese medicine, different herbs are used together for additive or synergistic effects.2The prevalence of Western herbal use among surgical patients ranges from 2220to 34%.21Not surprisingly, we found a high rate of recent TCHM use in our population (88%), reflecting the important role of traditional Chinese medicine in Chinese culture. The use of herbal remedies in Hong Kong is established and supported by an infrastructure of Chinese herbal medicine shops, registered practitioners, pharmacists, and schools of Chinese medicine.4However, the distinction has to be made between dietary use and true medicinal use of Chinese herbs, which involved a smaller proportion of patients (8%). Before admission, 6% of the patients had consulted a traditional Chinese medicine practitioner.
The spectrum of herbal medicines used in Hong Kong is different from that reported elsewhere.4The selection of TCHMs for herbal prescriptions also seems to be different from those selected for herbal teas, herbal soups, and prepacked over-the-counter preparations (table 2). For herbs with potent properties, these are used more often in TCHM prescriptions than in herbal teas and herbal soups. To date, several reviews1,5,7,22have highlighted the potential serious problems associated with herbal medicines during surgery, but these reviews have been based on case reports and limited pharmacodynamic and pharmacokinetic studies.
Because fresh ginger was commonly used in our population for cooking, it was difficult to estimate the risk of adverse events related to ginger. In our study, fresh ginger was not associated with prolonged INR or aPTT compared with patients who did not take ginger in the preoperative period. We are uncertain whether this reflects the safety of ginger per se , because ginger was not used in an isolated and concentrated form by patients in this study, or whether the study lacked power to detect an association. In a recent open-label, three-way crossover, randomized controlled trial of healthy volunteers, INR and platelet aggregation were not affected by (dry) ginger intake.23
The actions of specific ginsenosides derived from ginseng vary and, in many instances, give opposite pharmacologic effects.24Because there is concern about the irreversible platelet inhibition caused by ginseng, its use should be discontinued at least 7 days before surgery.6In our study, some patients continue to take ginseng up to the time of surgery. We found no association between ginseng and prolong preoperative INR/aPTT or any cases with excessive bleeding, but we noted that ginseng was not used in an isolated form. Alternatively, the lack of an association was due to low power in the study. A recent finding suggests that ginseng alone does not affect INR or platelet aggregation.25
No epidural or spinal hematoma was detected in 133 patients taking some form of TCHM who underwent regional anesthesia. This finding supports the view of the Second American Society of Regional Anesthesia Consensus Conference on Neuraxial Anesthesia and Antithrombotic Therapy, which states that “Herbal drugs, by themselves, appear to represent no added significant risk for the development of spinal hematoma in patients having epidural or spinal anesthesia.”26However, our study was underpowered to examine the increased risk of bleeding associated with the concurrent use of other medications affecting the clotting mechanism in patients taking TCHM with known anticoagulant effects undergoing regional anesthesia.
Surgery is a time of fluctuating hemodynamics, physiologic shifts, intense metabolic changes, and protein catabolism and anabolism.27In addition, the potential for drug interactions to occur in anesthetic practice is great due to the wide variety of agents a patient may receive during the perioperative period.28In each TCHM-by-prescription decoction, there is typically between 12 and 18 herbs.17The German Scientific Society of Traditional Chinese Medicine recommends discontinuing TCHM, especially if taken with Western drugs, before anesthesia.29In this study, 5% of patients were taking both Western medication and TCHM by prescription concurrently within the 2 weeks before surgery. In hypertensive patients, a potential TCHM–drug interaction was between β blocker, TCHM with known antihypertensive effects, and anesthesia, which may result in perioperative hypotension. However, we did not identify any “probable” cases of unanticipated intraoperative or postoperative hypotension in patients using concurrent TCHM by prescription with known antihypertensive effects and Western antihypertensive drugs.
If possible, all herbal intake by the patient should be identified as part of the preoperative anesthetic assessment. The assessment of potential drug interactions was difficult, and only a limited number of cases could be identified because just over half of our patients knew all or some of the TCHM ingredients included in their prescriptions. For anesthesiologists and surgeons to recognize adverse effects and drug interactions associated with TCHM, we believe that patients should be encouraged to obtain a copy of their prescription (with legible handwriting) during their visit to the traditional Chinese medicine practitioner for identification. For difficult cases, one may need to refer to a multidisciplinary team consisting of a pharmacist, a chemical pathologist, a scientific officer, and a physician for assistance.17
Currently, no official guidelines exist in Hong Kong regarding patients taking herbal medicines before anesthesia,4mainly because of lack of reliable data. Using recent methodology30applied to our study, we estimate that if TCHM-by-prescription use was discouraged before surgery, 130 (95% CI, 16–390) preoperative events could be prevented at our hospital each year (assuming 10,000 procedures/yr). Although the level of knowledge among anesthesiologists in Hong Kong about TCHM and its adverse effects in surgical patients is unknown, a recent US survey has highlighted the need for educational interventions to improve on the lack of knowledge about adverse effects of common herbs among healthcare professionals.31This study also highlights the need for patient education materials to be developed to increase the safe use of TCHM.
There are several limitations to this study. Because there was a high prevalence of TCHM use before surgery in this population, we acknowledge that the non-TCHM control group was small and may have led to low power in this study. Stopping TCHM within 2 weeks of surgery may not be always feasible because of the patient's strong belief in the efficacy of traditional Chinese medicine.4However, we believe that the true medicinal use of TCHM is measured by the patient's prescription use rather than by self-prescribed use (mainly for health promotion purposes). Therefore, we examined the association between TCHM-by-prescription use and adverse perioperative effects. We relied on patients' recollection of their history of recent TCHM intake. The reliability of these data are unknown without an inspection of herbal residues to confirm that only the intended herbs have been taken and screening of herbal products and biologic samples for the presence of synthetic drugs and specific toxic ingredients.17In this study, we focused on the adverse effects of TCHM itself rather than on the effects resulting from failure of good manufacturing, dispensing, or prescribing practice. However, the latter may have been involved in some of the preoperative events identified. The difficulties and challenges of initiating and completing an outcome-based herbal study in anesthesia cannot be underestimated. In conducting future outcome-based TCHM studies, complex issues of quality control, adulterations, and synergy between different TCHMs must be addressed.
In summary, our study provides additional evidence of the need for healthcare professionals to be aware of the potential preoperative risk associated with TCHM-by-prescription use before surgery. Health professionals should also inquire about the patient's history of TCHM use during the preoperative assessment. Patients should be educated to stop using TCHM by prescription near the time of surgery.