Hay Fever Treatment in General Practice: A Randomised Controlled Trial Comparing Standardised Western Acupuncture with Sham Acup

Hay Fever Treatment in General Practice: A Randomised Controlled Trial Comparing Standardised Western Acupuncture with Sham Acupuncture


The effect of standardised, Western acupuncture on hay fever symptoms was investigated in a randomised, controlled, single-blind trial in comparison with "sham" acupuncture. Three general practices, in Oxfordshire (rural), Lincolnshire (semi-rural), and Peterborough (urban), recruited 102 patients aged 16 or over with long-standing, moderate or severe hay fever symptoms that had required continuous therapy for at least one month of the year for three or more consecutive years.

The patients were asked to keep a diary to record: the amount of medication used daily; a daily symptom score (using a ten-point scale), from which was derived a weekly remission of symptoms score; and their assessment of the effect of acupuncture on the hay fever symptoms. Symptom scores and use of medication were similar in the two groups. In the four-week period following each patient's first treatment, remission of symptoms was reported by 39.0% in the active treatment group and 45.2% in the sham group; mean weekly symptom scores were 18.4 and 17.6 respectively; and mean units of medication used were 4.1 and 5.0 respectively. Sixteen out of 43 patients in the active treatment group and 14 out of 43 in the sham group felt that the acupuncture had had an excellent or very good effect on their hay fever. The treatments were simple, safe, reproducible and perceived as equally effective. Whether this represented an acupuncture effect, a placebo effect, or natural variation in a fluctuating condition, is not clear.

Key words

Acupuncture, General practice, Hay fever, Randomised controlled trial.


Hay fever is a widespread and increasing problem affecting up to one third of the young adult population during early and mid summer ( 1-6). Those patients whose symptoms are suppressed by medication often suffer from drug side-effects and still feel and perform below par ( 7). Traditional acupuncture has been used for many centuries to treat hay fever symptoms ( 8, 9). Patients are usually given an individual prescription of points which may alter with each treatment, but this form of therapy is time-consuming and does not lend itself well to general practice. It is particularly difficult to study systematically because of the considerable interpractitioner variation in the way treatment is given ( 10, 11).

In a pilot study, Williamson ( 12) showed it was possible to offer hay fever prophylaxis in general practice using a single point, minimal technique. In this pilot study of 31 patients with long-standing, moderate or severe hay fever, those treated with acupuncture had fewer symptoms during the hay fever season and a greater improvement over the previous year than untreated controls. The aim of the present study was to test the effectiveness of this form of Western acupuncture in improving hay fever symptoms, using a larger number of patients in three general practices.

Acupuncturists can treat hay fever symptoms either by giving prophylaxis during March or April, or by treating symptoms as soon as they start, or both. In the present study, hay fever treatment rather than prophylaxis was studied. This approach avoids treating patients whose symptoms would have resolved spontaneously ( 1).

The points used for active treatment in the study were chosen after careful discussion between the three participating doctors, who were experienced general practitioner acupuncturists, and with the scientific committee of the British Medical Acupuncture Society. No form of acupuncture constitutes a genuine placebo treatment, since a needle anywhere in the skin causes neuropeptide release and is therefore a form of acupuncture ( 10, 11, 13). However in a previous, randomised, controlled trial of acupuncture for treating disabling breathlessness ( 14), the control treatment was sham acupuncture using "dead" points along the medial border of the patella. Although all patients improved, there was a greater effect with the traditional Chinese acupuncture than with the sham acupuncture. Following this example, we used anterior knee points as our sham treatment.


Ethical approval for a randomised, controlled, single-blind trial of standardised, Western acupuncture compared with sham acupuncture was granted by the Central Oxford Research Ethics Committee.


The patients came from three general practices: in Faringdon, Oxfordshire (rural), Stamford, Lincolnshire (semi-rural) and Peterborough, Cambridgeshire (urban). The general practitioners involved all had at least five years of acupuncture experience. Patients were eligible for the trial if they were aged 16 or over, and had had moderate to severe hay fever symptoms, requiring continuous therapy for at least one month of the year, for three or more consecutive years. Pregnant women were excluded.

Patients were recruited before the start of the hay fever season in 1994, by letter, surgery poster, and local advertisement. Responding patients were asked to complete a screening questionnaire to ensure their eligibility for the trial, and to ascertain the severity of their hay fever symptoms in 1993. All those eligible were requested to attend as soon as symptoms began. Those who presented with symptoms gave informed consent and were randomly allocated, in blocks of four, to the two treatment groups. The random allocations were placed in numbered, sealed, opaque envelopes and were opened by the general practitioner just before treatment was started. Patients were told that they were taking part in a trial to test whether acupuncture could help their hay fever symptoms. They were told that there were two treatment groups comparing different forms of standardised acupuncture.


Patients recorded the severity of their symptoms using a ten-point scale. Before entry to the trial, they made an overall assessment of symptom severity during the 1993 season. At the end of the trial, they made a similar assessment for 1994, as well as recording side effects and opinions about the acceptability and efficacy of their treatment. During the 1994 season, they were asked to record the severity of their symptoms each day, and to make an overall assessment of symptom severity each week. Remission of symptoms in any week was defined, before the start of the trial, as a total weekly symptom score of 14 or less (maximum 70).

Patients were advised to use whatever medication they needed and to record the type and dose used each day during the 1994 season. Medication was classified by site of use: systemic (antihistamine), and local treatment to the nose, eyes and chest. One unit of medication was defined as the use of any amount of medication, at any one site, on one day. Patients could therefore use a maximum of four units per day.

The recordings were analysed for the four weeks starting 6th June 1994 (the first week in which hay fever symptoms became widespread) and for the four weeks following each patient's first treatment. The study was designed to detect, with à = 0.05 (two-tailed test) and power 0.90, an increase in the proportion of patients showing remission of symptoms from 20% in the sham acupuncture group to 50% in the Western acupuncture group. This required 58 patients in each group.


Special hay fever clinics were run weekly in the participating practices from the last week in April until the second week in July. Patients were instructed to attend the clinic during the week that their hay fever symptoms started and were offered three or four acupuncture treatments at weekly intervals. Three acupoints were treated bilaterally in all patients, using 32 gauge, disposable needles.

1. Active treatment

Bladder 2 (Zanzhu): at the medial end of the eyebrow, directly above the inner canthus of the eye to a depth of 3mm.

Large Intestine 20 (Ying Xiang): between the ala nasi and the nasolabial groove to a depth of 2mm.

Large Intestine 4 (Hegu): the highest point of the adductor pollicis with the thumb and index finger adducted to a depth of 1-2cm.

Immediately on insertion, needles were stimulated to obtain the de qi sensation, left in situ for five minutes with no further stimulation applied, then removed.

2. Sham treatment:

Three needles were inserted just subcutaneously, in an area over the middle of the patella, to a depth of 1-2mm. Needles were left in situ for five minutes and then removed with no further stimulation.

In both groups, the contact with the doctor was minimal. The needles were simply inserted and the patients left to lie quietly for five minutes, when the doctor returned and removed the needles.

Statistical analysis

Results are presented for each week of the four-week periods analysed (Tables 2, 3 and 4), but the numbers of observations vary because of missing data. The sum of the symptom scores plus units of medication used in each four-week period was calculated for every subject who had supplied data relating to at least one week of that period. These figures were then used as a base for all hypothesis tests relating to the four-week periods. More than 80% of these subjects had complete data for the four weeks. Remission of symptoms for a four-week period had been defined before the start of the study as a mean weekly symptom score of 14 or less.

Numerical variables were compared using the t test (and the results confirmed with a non-parametric test), proportions were compared using the X( 2) test, and comparisons within individuals were made with paired tests.


One hundred and sixteen eligible patients were recruited, of whom 102 presented for treatment (Table 1), and were randomly allocated to the active or sham groups (51 in each). Patient recordings were inadequate or missing for 13 of the 102 treated patients. Of the 89 remaining subjects, 86 (84.3% of those treated) completed questionnaires relating to both 1993 and 1994, and 84 (82.4%) completed daily diaries.

The mean (SD) age was 31.9 (10.2) years in the active treatment group and 29.9 (7.5) years in the sham group. The proportions of males were 19 out of 51 (37.3%) and 23 out of 51 (45.1%) respectively.

Symptoms and medication

During the 1994 season, symptom scores and use of medication were very similar in the two treatment groups, whether analysed for the four-week period starting 6th June 1994 or the four-week period following each subject's first treatment (Tables 2,3 and 4). In this latter period, remission of symptoms was reported by 39.0% in the active treatment group and 45.2% in the sham group (difference 6.2%; 95% confidence interval -27.4 to 15.0); the mean (SD) weekly symptom scores were 18.4 (11.2) and 17.6 (11.2) respectively (difference 0.8; 95% confidence interval -4.1 to 5.7); and mean (SD) weekly units of medication used were 4.1 (4.6) and 5.0 (5.0) respectively (difference -0.9; 95% confidence interval -3.0 to 1.2).

At the end of the 1994 season, 16 out of 43 (37.2%) in the active treatment group, and 14 out of 43 (32.6%) in the sham group, reported that they felt an excellent or very good response to the acupuncture (difference 4.7%; 95% confidence interval -15.5 to 24.8) (Table 5); this was defined as having negligible symptoms, requiring no or reduced medication. A further 16 out of 43 (37.2%) in each group felt they had had a good result from the acupuncture (fewer symptoms than usual, with no more medication).

Perceived overall symptom severity was reduced to the same extent in both groups from 1993 to 1994 (Table 5). Using a ten-point scale, the mean reduction in the active treatment group was 2.0 (95% confidence interval 1.2 to 2.8; p<0.001) and 1.5 (95% confidence interval 0.8 to 2.2; p<0.001) in the sham group.

Acceptability of treatment

All patients reported that they either enjoyed or could tolerate the treatment and the majority said that they would definitely consider having acupuncture again for their hay fever: 31 out of 44 (70.5%) in the active treatment group compared with 26 out of 44 (59.1%) in the sham group (difference 11.4%; 95% confidence interval -8.5 to 31.2). Only two patients in each group said they would not consider having acupuncture again.

Few side effects were reported. Moderate or severe pain on insertion of needles was reported by 9 out of 44 (20.5%) in the active treatment group and 4 out of 43 (9.3%) in the sham group (difference 11.2%; 95% confidence interval -3.6 to 25.9), and moderate or severe pain while the needles were in' situ by 2 out of 44 and 2 out of 43 patients respectively. Minor bruising at the acupuncture site occurred in 9 out of 44 (20.5%) of patients in the active treatment group and 8 out of 43 (18.6%) in the sham group.

Pollen count

Daily pollen counts for the months of June and July were obtained from the Environmental Health Departments in Oxford and Lincoln. Median (range) counts in Oxford were 34 (1 to 186) in 1993 and 52 (0 to 286) in 1994. Comparable counts for Lincoln were 19 (0 to 96) and 18 (2 to 120).


Not as many patients as intended were included in the trial. Fourteen of the 116 eligible patients did not present for treatment for various reasons, including resolution of symptoms. In addition, there were some missing recordings, as not all treated patients, returned their questionnaires and diaries. However, we have no reason to suppose that people who had not returned their recordings were any different in the two groups.

No difference in outcome was observed between the two treatment groups. This contrasts with the trial reported by Jobst et al. ( 14) of 26 patients with chronic obstructive pulmonary disease, in which there was a statistically significant benefit in the group treated with "true" acupuncture. However there are important differences between the studies. The patients in this study were healthy volunteers with a self-limiting illness, and the acupuncture was standardised and given in a busy general practice setting. Those in the Jobst et al. study had end-stage respiratory disease, and were treated with traditional Chinese acupuncture, given in a relaxed hospital setting, with no time constraints.

The two groups in our study were similar, not only in the severity of their symptoms during the 1994 hay fever season, but also in their perceived improvement over the previous season. This result can be interpreted in three ways:

i. Both treatments could have been effective. Mann has observed that many patients respond better to lesser, minimal needling than to conventional, more painful acupuncture ( 10). Lewith and Vincent have further hypothesised that in conditions such as addiction, where the effect of acupuncture is largely mediated by endorphins, the number of needles rather than the site is of most relevance ( 13). It is unknown whether hay fever is one of these conditions.

ii. The second possibility is that both sham and active acupuncture are more effective than no treatment, but both are working as placebos. This would suggest that the simpler method of acupuncture to carry out, that is the sham treatment, would be preferable, especially as the site is very safe and the patients tend to experience less pain.

iii. Thirdly it could be that neither treatment was effective and the milder symptoms reported in 1994 compared with 1993 were due to faulty recall, or to random variation. Ross and Fleming ( 1) have reported that hay fever symptoms tend to increase with the pollen count, and it was notable that the pollen count was slightly higher in 1994 than 1993, so that an increase rather than a decrease in symptoms might have been expected. The extent of random variation could only have been determined by including a third group in the trial, who received no acupuncture at all. Although Vincent and Lewith extensively discussed the validity of placebo controls for acupuncture studies, they failed to recognise the importance of including an untreated control group in studies of conditions that vary within a patient over time ( 15).

Were the acupuncture treatments shown to be clinically effective, then they would be cost effective. The material costs were very small (36p per treatment). The doctor's time with each patient could have been fitted into four standard surgery appointments and, given the standardisation of the treatment, there is no reason why it could not have been carried out by a practice nurse. In addition to the relief of symptoms, the low cost to the patient was beneficial, since usual hay fever treatment is bought over the counter or incurs prescription charges. Most patients suffered only minimal discomfort, and 95% of patients were willing to consider acupuncture again.

1. Ross AM, Fleming DM (1994) Incidence of allergic rhinitis in general practice, 1981-82. British Medical Journal. 308: 897-900

2. Richard S, Thornhill D, Roberts H, Harrier U (1992) How many people think they have hay fever and what they do about it. British Journal of General Practice. 42:284-6

3. Fleming DM, Combie DL (1987) Prevalence of asthma and hay fever in England and Wales. British Medical Journal 294:279-83

4. Pedersen PA, Weeke ER (1981) Allergic rhinitis in Danish general practice. Allergy. 36: 375-9.

5. Ewan PW (1989) Hay fever: Report of the Royal College of Physicians. Journal of the Royal College of Physicians of London. 23:757-62

6. (1986) Hay fever among London's taxi-drivers (Letter). Lancet. 2:1397

7. Vuurman EFPM, van Veggel LMA, Uiterwikj MMC, Leutner D, O'Hanlon JF (1993) Seasonal allergic rhinitis and antihistamine effects on children's learning. Annals of Allergy 71: 12-6

8. Xinnong C (1990) Chinese Acupuncture and Moxibustion. Foreign Languages Press, Beijing

9. Veith I (trans) (1966) The Yellow Emperor's Classic of Internal Medicine. University of California Press, Berkeley

10. Mann F (1992) Reinventing Acupuncture. A New Concept of Ancient Medicine. Butterworth Heinemann, Oxford

11. Vincent CA (1993) Acupuncture as a Treatment for Chronic Pain. In: Lewith GT, Aldridge DA eds. Clinical Research Methodology. Hodder and Stoughton, London: 289-308

12. Williamson L (1994) Hay fever prophylaxis using single point acupuncture: a pilot study. Acupuncture in Medicine 12(2):84-7

13. Lewith G, Vincent C (1995) On the evaluation of the clinical effects of acupuncture: a problem reassessed and a framework for future research. American Pain Society Journal. 4(1): 29-39.

14. Jobst K, Chen JH, McPherson K, Arrowsmith J, Brown V. Efthimou J, Fletcher H, Maciocia G, Mole P, Shifrin K, Lane D (1986) Controlled trial of acupuncture for disabling breathlessness. Lancet. 2:1416-8

15. Vincent C, Lewith G (1995) Placebo controls for acupuncture studies. Journal of the Royal Society of Medicine. 88:199-202

The British Medical Acupuncture Society.


By Patricia Yudkin; Ruth Livingstone; Koneru Prasad; Alice Fuller and Martin Lawrence

Share this with your friends