Consent for a randomised-groups design, case controlled, open food challenge, using the E&RP, cross-over study on adult asthmatics, was obtained from the Ethics Committee of the South African Medical Association (SAMA). The same protocol was accepted by the ethical committee of the University of the Witwatersrand (RAND)

Asthmatic adult volunteers, recruited by media advertisements, all underwent the initial screening requirements of the trial.

Volunteers were accepted into the trial when the Forced Expiratory Volume in 1 second (FEV1), measured on a spirometer (Micro Loop, Micro Medicals, Kent, England) satisfied the conditions of the protocol, which were:

a) If the initial FEV1 was less than 75% of predicted value.

b) or if the initial was greater than 75%, a fall of more than 15% of the FEV1 predicted value, after exercise using Jones method ₂₂ and returned to normal after salbutamol inhalation.

or after exercise or if too distressed to exercise, the FEV1 predicted value improved by 15% or more following salbutamol inhalation.

When the above requirements were met the volunteers were randomised by JML into Control (C) and Diet (D) groups, according to a statistically evaluated, randomising table. A flowgram of lung function, using the FEV1 as the standard criterion was measured:-

Figure 1

The incidence of nasal symptoms and improvement, amongst these asthmatics, such as sneezing, itching, pre and post nasal drip, congestion and throat clearing was recorded. The improvement of the above symptoms was recorded daily after each meal and noted at each weekly visit for the six weeks of the trial and follow up periods.

Each volunteer was informed to which group he/she had been randomised and the purpose of the particular group was explained. Each volunteer completed a questionnaire detailing severity and frequency of symptoms of rhinitis and asthma and dosage of current medication. Informed consent was signed by all. The questionnaire was again completed after 6 weeks, the period of the trial and again after 6 weeks of the cross-over period.

This group was seen weekly for 6 weeks when the lung function (FEV1), current symptoms and the dosage and frequency of medication used, were recorded. They were informed to carry on with their usual therapy. Only after 6 weeks the control group was given the information and option to go onto the E&RP. Their symptoms and medications during the two periods were compared, with each person then acting as his or her own control.

Each patient completed an extra food questionnaire detailing their usual diet, and if any foods had affected him/her in any way. The principle and technique of the E&RP was explained to each volunteer and given a menu on a diet sheet, and instructions detailing the requirements of the diet.

The diet involved eating one or two different types of foods in their natural state at each meal. ₂₁

Each day a different food of the food types were to be eaten, such as grains for breakfast, fruits for lunch and a protein with a vegetable for supper. Foods were to be bought fresh or frozen and eaten raw or either cooked, steamed, baked, micro-waved or grilled but not fried. No processed foods (bread, pizzas, confections, sauces, cold drinks, coffee) were allowed. During the first week only water was allowed to be drunk and salt the only condiment added. The specific diet was prescribed on a tabulated menu-form, according to the likes and dislikes of each individual, with the food to be eaten at each meal for each day of the week for the 6-week period. No food item being repeated in the week. ₂₁(Figure 1)

Figure 1: The elimination diet menu sheet.

The diet sheet shows the menu with one or two different types of foods to be eaten, in the natural state, in unlimited amounts to satisfy the appetite, at each meal for the week No food item is repeated in the week. After each meal, symptoms are entered, using abbreviations, as listed and graded according to severity: 0-4:

O=no symptoms. 1=mild. 2=fair, no medicines needed. 3=needed medicines. 4=severe, confined to bed.

The symptoms were relieved within three days and provoked after eating green apples, for breakfast on Saturday. The second week another fruit would replace the apples. If no improvement, then the third week pork chops and spinach of te supper the night before would be replaced. If still no improvement, then in subsequent weeks, the food/s of previous meal would be replaced, until symptoms are relieved. This method identifies the time it took for the offending food/s to trigger the symptom/s. Medicines and dosage are also recorded.

Name_______ ___ ____________Age____Gender M/F. Smoke Y/N. Complaint__________Problems: Normal hours 012-345-6789. Fax 012-123-4567

Grading of symptoms: Severe in bed -4. Had to take medicine -3. Bearable -2. Fair -1. No symptoms -0. After hours: Phone and Fax 012-123-4568

Date: / / Week No. Weight__________Kgms

Figure 2

After each meal, any symptoms present were recorded and graded by the patient, as to severity, as well and dose of medication needed. Any food that provoked symptoms was replaced by another food, the following week, until the patient was symptom free. Patients were seen weekly and any further queries regarding the requirements of the diet were addressed. Over the six weeks other beverages, condiments and other foods, including processed foods were added and if any provoked symptoms were replaced. After 6 weeks those who were not completely relieved of symptoms then had inhalants investigated, to ascertain the role any inhalant played in each volunteer.

It must be noted that all patients were ambulant. One patient was on a monthly injection of cortisone, treated by his physician, 40 were on inhaled and 4 on oral corticosteroid therapy. All were on various long and short acting bronchial dilatation therapy. As asthma decreased or disappeared, both long and short acting bronchodilators were reduced or stopped completely. Those on inhaled corticosteroid therapy were tapered gradually. The dosage and frequency of medicines with the degree of relief were also recorded, on the diet sheet. Patients were requested to avoid any bronchodilator medication, on the morning of exercise testing. The follow up period was to be at three monthly intervals for one or two years, if possible.

The mean change in the FEV1 predicted va1ues between the control and trial groups, pre- and post- trial, was calculated using a paired Student-t test.

Of the 78 patients who completed the diet programme, foods alone provoked episodes of wheezing and cough in 49 (63%), foods combined with inhalants in 21 (27%) and 6 (8%) of the 8 (10%) who did not respond to the diet, improved after avoiding inhalants. In two patients (2%) no trigger, neither food nor inhalant, was identified.

Fruits were identified as the triggers in 83 (29%) instances, followed by proteins in 74 (26%), grains, in 61(21%), vegetables in 35 (12%), nuts in 5 (2%) and processed foods in only 27 (10%). Contrary to expectation, preservatives, colourants and so-called “processed foods” played a minor role. Chocolate was the trigger in 6 instances, wine in 4, fruit juices in 3 (orange juice in 2, guava juice in 1), while, ice cream, honey, coffee and tea and beer, accounted for 2 patients each. Custard cake, guava roll, tomato sauce, chicken and beef burger affected only one person each. Four patients, who had adhered to the diet for the 6 weeks, rewarded themselves by indulging in their favourite food (honey in 2, and a chicken burger and beefburger in one each) and reported an immediate onset of asthma. Some patients reacted to only one food, while others from two six, different foods. (TABLE 2)

Figure 4

Table 2: The incidence of the various foods that provoked the symptoms of asthma as identified by the elimination and rotation diet and subsequent avoidance over a period of six weeks and follow up of 24 -36 months.

Nineteen of the 78 patients, not adequately improved after 6 weeks on the diet, then had inhalants investigated. Perfumes, soaps, shampoos, pets, cigarette smoke, exhaust fumes, work situations, shopping centres, industrial sites, time of day and season of year were eliminated where applicable. The only laboratory test used, was when symptoms were related to being at home, to exclude house dust mite sensitivity. Inhalant factors included, cigarette smoke in 12 instances, (15%) (8 passive and 4 active), house mite and dog hair in 4 (8%) each, cat hair in 3 (3%), perfumes and sprays in 2 (2%) each. Printers ink, car exhausts, damp walls due to a leaking roof, grass cuttings, saw dust, industrial area, smog, rain storm and grass fire affected one patient each. Multiple factors were involved in those reacting to both foods and inhalants.

Nasal symptoms, such as sneezing, itching, excess mucous and nasal congestion occurred in 77 (75%) of the 102 asthmatics. Amongst the control group these symptoms occurred in 36 (70%), Only 2 improved when given increased doses of corticosteroid nasal sprays.

Amongst the 78 patients who completed the diet programme, avoiding the offending foods, nasal symptoms present in 64 (82%), cleared in 43 (67%), improved in 17 (27%) while 4 showed no change. Los of sense of smell, taste and hearing also improved. Of the 4 (6%) who had not improved 3 were subsequently shown to react to dogs, cats and cigarette smoke. In one no cause was found. (Table 3).

Figure 5

Table 3: The incidence of sinusitis in the subjects of the present study before and after the 6 week trial period. Changes in the sensory modalities of taste, smell and hearing are shown as well. Tinnitus in 1 was relieved. The sight in one improved remarkably.

Smoking played a minor role. Of the 102 asthmatics more than half, 69 (68%) hade never smoked. While 21 (20%) had stopped smoking between 2 and 27 years previously. Only 12 still smoked. In patients with FEV1 predicted values less than 50% 9 had never smoked and 7 had stopped smoking more than 2 years peviously and only one an active smoker. In 8 patients passive cigarette smoke was the trigger of their wheezing. ( Table 4)

Figure 6

Table 4: The incidence of smoking in this trial was surprisingly low. Of the 102 asthmatics more than half, 69 (68%) had never smoked while 21 (20%) had stopped smoking between 2 and 27 years previously and 12 (12%) still smoked. In the 17 patients with an FEV1 of less than 50%, 9 had never smoked while 7 had stopped smoking more than 2 years previously and only 1 was an active smoker. In eight patients passive cigarette smoke was the trigger of their wheezing. (TABLE 3).

As symptoms were relieved, patients on inhaled long acting bronchodilators were asked to reduce the dosage from two puffs twice a day (bd) to one puff bd, then once daily and tapered off to alternate days, every second or third day and ceased when free of symptoms for four days. Similarly for short acting inhaled bronchodilators and then only when necessary. When free of bronchodilator therapy for a week, oral or inhaled corticosteroid therapy, was tapered off, over three to four weeks. Progress in reduction of medication, was noted at each visit and follow up, for one or more years.

Of the 78 patients who were on various inhaled bronchodilator medications only 17 needed such therapy occasionally, when they inadvertently ate a processed food, containing the trigger or when they wanted to test the identified food, their favourite. (Table 5) Similarly of the 67 patients on oral or inhaled corticosteroids only 8 needed this therapy after avoiding the trigger foods. Four of them though free of symptoms were advised by their physicians not to stop this therapy. (Table 6)

Figure 7

Table 5: Indicating the doses and number of patients on inhaled bronchodilators before and after the E&RP. After avoiding the offending food/s, the 16 patients who occasionally needed a bronchodilator, are those who advertently ate a processed food that contained the trigger of their symptoms, not previously identified, during the follow up period.

Figure 8

Table 6: This table shows the incidence of oral and inhaled corticosteroids used by the patients before and after the elimination diet. One patient who took 12 tablets of prednisone, 5 mg each daily, only needed 1 tablet qid. Of the 18 patients on oral cortisone therapy only two needed reduced dosages. Of the 49 patients on inhaled corticosteroid therapy before the diet, 43 were relieved of their symptoms and only 6 were on reduced doses, after the diet.

The one patient who received cortisone by injection monthly did not need further injections.

Nebuliser inhalation therapy used by 10 patients became redundant in all but 2. Five on fenotrol 1.25 nocte, 4 on Ipatropium bromide and 1 on Fenotrol/Ipatropium were free of therapy except 2 who needed fenoterol prn. Sodium cromoglycate, by inhalation, was used by two.

On follow up after 6 months, though improved and free of symptoms, four of the above were advised by their physician s not to stop taking their previous cortisone medication, to which they complied.

Antihistamine use by 8 on Loratidine 10 mg daily, 3 on Cetrizine HCL 6 mg daily, 1 on Ketotofin 1 mg bd and another on Oxatomide 30 mg bd, including 2 patients on Singulair 10 mg inhalations, was unnecessary after the diet.

By avoiding foods identified by the E&RP, a statistical improvement of 21.00% of the predicted FEV1 value in the diet group compared to controls who were on their previous medication ( p<0.0001). Forty nine (63%) of the 78 asthmatics in the diet group, discontinued all medications, while 21 (27%) were taking less after six weeks. The diet did not help 8 (10%), 6 (8%) of whom afterwards, reacted to inhalants. Only two patients did not benefit.

After two to three years, 66 of the original 78 persons on the diet protocol, were traced. Patients were encouraged to return to normal diets but to exclude offending foods. Forty nine (74%) patients were symptom free, without medication, 9 patients (14%) needed bronchodilators, when inadvertently they had eaten offending food(s). Five patients (8%) could not resist eating identified offending foods, (favourite foods), inhaling bronchodilators before eating them. Two patients (3%) were unwilling to forego their favourite food, (beer and toasted cheese) and resumed medication. One would not remove her pet cat and one would not stop his woodwork hobby. Two (3%) pensioners could only afford bread and cheese, their triggers and required medication. Four (6%) patients, though symptom free, were advised by their physicians to continue with their corticosteroid therapy, which they did.

During the follow-up of two to three years two patients, who had recorded the lowest FEV1 readings of 12 and 16%, had died of causes unrelated to asthma. Their quality of life, prior to death, had improved because of freedom from asthma symptoms, aided by reduced doses of inhaled corticosteroids. Generally absenteeism was much less, they were able to do much more work, without medication, but some remained on reduced doses of inhaled corticosteroids.

The trial was made possible by a grant from the BEARE Foundation for medical research, Durban, South Africa.