Brittle asthma: Management

Patients with brittle asthma pose difficult and complex management problems. Jon Ayres identifies possible management strategies for these patients.

Jon Ayres, BSc, MD, FRCP, is Professor of Respiratory Medicine (University of Warwick) in the Department of Respiratory Medicine, Birmingham Heartlands Hospital, Birmingham.

The term ‘brittle asthma’ was first used by Turner-Warwick (1977) to describe patients with asthma whose peak expiratory flow (PEF) varied ‘chaotically’; a pattern which could lead to death from an acute severe, attack (Bateman and Clarke, 1979; Westerman et al, 1979). Morbidity from brittle asthma is considerable, and this review aims to identify possible causes and management strategies for this group of patients.

Since 1977, the term brittle asthma has been used in different ways by different physicians, leading to some confusion over whether such a group is truly separable from other patients at the severe end of the asthma spectrum. In order to try and clarify this area we have suggested a classification of brittle asthma into two types; a feature of both types being a susceptibility to repeated severe attacks resulting in hospital admission.

Type 1
Patients who consistently demonstrate wide peak flow variation (greater than 40% diurnal variation for at least 50% of days), despite maximal medical therapy including at least 1500 µg/day of inhaled beclomethasone or equivalent, are classified as having type 1 brittle asthma.

It is crucial that peak flow readings are corrected for non-linearity (Miller et al, 1992) when calculating diurnal variation in type 1 brittle asthma. These patients are typically female (4:1 in our clinic) and aged between 15 and 55 years.

Type 2
Patients with type 2 brittle asthma appear to be well controlled between attacks which are often sudden in onset (occurring within minutes) and are associated with loss of or disturbed consciousness on at least one occasion.

The patient may not have been mechanically ventilated as a result of the attack but is very likely to exhibit a severe respiratory acidosis in an attack and, if ventilated, needs support for relatively short time periods (Wasserfallen et al, 1990). Type 2 brittle asthma seems to be equally prevalent in men and women.


Type 1 brittle asthma is a cause of significant morbidity associated with frequent accident and emergency attendances and hospital admissions and the resultant use of considerable amounts of medication. Consequently, side effects of therapy, particularly oral corticosteroids, are common, e.g. osteoporosis, weight gain and oesophageal reflux (Miles et al, 1993). In addition, this can result in the development of obstructive sleep apnoea which may remain unsuspected because poor sleep quality tends to be attributed to asthma, even though the symptom pattern is exactly the same in these patients compared to isolated sleep apnoea. Within our clinic 12% of subjects suffer from sleep apnoea and are treated with nasal continuous positive airway pressure.

Risk factors

Over 90% of patients with type I brittle asthma are strongly atopic (Miles et al, 1995), despite which over two-thirds keep pets at home, so continuing exposure to aeroallergens may be important in maintaining symptoms in this group, especially as they spend more time at home and indoors.

Food intolerance
Over 60 % of type I patients report at least one food or drink which makes their asthma worse. Double blind placebo-controlled food challenge (Baker et al, 1996) has confirmed a prevalence of food allergy at over 50% in type I brittle asthma with wheat and dairy products being the most important triggers (Table 1).

Table 1: Positive Food Challenges in Brittle Asthma

Food % Positive Response
Dairy Products 50
Wheat 50
Fish 37
Citrus 34
Egg 34
Potato 26
Soya 23
Peanut 16
Yeast 13

Psychosocial factors
Psychosocial factors are important in type I brittle asthma (Garden and Ayres, 1993), with a high incidence of depression and frequent evidence of broken relationships and physical and sexual abuse in our experience. Just as common are abnormal coping strategies for managing deteriorating asthma, where panic frequently supervenes (Miles et al, 1997). However, it is difficult to be certain whether brittle asthma is associated primarily with personality disorder, or whether the threat of severe asthma induces psychological instability.

Steroid responsiveness
Patients with type I brittle asthma are usually treated with high doses of inhaled/oral steroids, yet their asthma often remains poorly controlled suggesting that there may be a degree of resistance to the anti-inflammatory effects of steroids (Barnes and Adcock. 1995).

Management: Type I brittle asthma

These patients are, by definition, extremely difficult to manage. Many of them have fallen out with their doctor, who perhaps understandably. has run out of ideas and, often, patience. Management should be holistic, trying to approach all areas which impact on an individual’s symptoms. This involves identifying causal factors and dealing where possible with psychosocial factors, before attempting to wrestle with polypharmacy. Trying to identify psychosocial factors and dealing with them (including help with Disability Living Allowance, for instance) can help significantly, and group therapy has been of some use in patients who are able to meet on a regular basis, by providing an auto-support network.

Allergen exposure
Removal of animals and determined efforts to reduce house dust mite exposure would seem logical but there is no evidence to suggest that such control measures are effective in this group. Indeed, given their psychosocial problems, these patients invariably believe that their animals are more reliable than people, and suggesting that their animals should be removed is invariably greeted with blunt refusal.

Where foods are identified as allergic triggers they should be avoided. In some cases such avoidance can be remarkably effective while in others the benefit is limited, but compliance with what is often a difficult diet may not always be good. Good dietary advice and support is crucial as Baker and colleagues (1995) have shown that the diets of these patients are very often deficient in minerals such as selenium and magnesium and in the anti-oxidant vitamins A, C and B. This is probably because their diet lacks food that they are avoiding because of intolerance and a desperate need to lose weight with consequent reduction in calorie intake.

The most important factor to identify is compliance with treatment but it should be remembered that these patients appear to have wide variations in peak flow and symptoms before they opt not to take treatment as we would wish. Non-compliance is not a cause of the severe asthma itself. Where identified this should be openly addressed with the patient and their carer.

Drug therapy
The BIS guidelines (British Thoracic Society et al, 1996) are of limited help in these patients. They are already taking large doses of inhaled steroids leaving an increase in, or resumption of oral steroids as the next therapeutic step when symptoms deteriorate, which many resist because of side effects. Consequently, many patients simply increase their beta-2 (1-2) agonist use (Miles et al, 1997) in an attempt to avoid oral steroids and hospital admission. Whether alternative immunomodulatory treatment, such as methotrexate or cyclosporin, will be effective is not yet certain but is worth trying in individual cases.

Subcutaneous §2 agonists
Type I patients with brittle asthma can be treated with long-term continuous subcutaneous infusion of §2 agonist, usually terbutaline (CSIT) (O’Driscoll et al, 1988; Ayres, 1992). Using this technique, around half of patients with type I brittle asthma show marked improvements in symptoms, variation in PEF (Figure 1) and use of other asthma medication, including oral steroid use. Around 25 % show some improvement in symptoms but less improvement in PEF while the remainder do not respond. Chronic steroid-dependent asthmatics without intrinsic PEF variability do not respond to this form of therapy.

Figure 1: Peak flow readings (before and after bronchodilator) in a patient with type 1 brittle asthma

Before Subcutaneous Terbutaline
Salbutamol Inh.x n
Prednisolone 5mg
Salbutamol SR 8mg b.d.
Uniphylline 800mg nocte
Pulmicort ii q.d.s
Intal 5 iii q.d.s

On Continuous Subcutaneous Terbutaline
Salbutamol ii p.m.
Pulmicort ii q.d.s
Intal 5 ii q.d.s

On Terbutaline subcutaneously in 4 daily divided doses.
Salbutamol ii p.m.
Pulmicort ii q.d.s
Intal 5 ii q.d.s

The infusions are given through a battery-powered syringe driver (such as the Graseby MS26, Graseby Ltd, UK) which need to be provided in successful cases on a long-term basis. If provision of a pump is not possible then the daily dose can be delivered by divided doses, albeit with slightly less effective control (Figure 1). The best tolerated subcutaneous needles are the Sof-Set and the Disetronics needles (Applied Medical Technology, Cambridge, UK), which although more expensive than the standard butterfly, last much longer; one needle often remains in situ for a week or more compared to 48 hours or less for the butterfly needle.

The usual dose needed ranges between 6 and 15mg a day, mean blood levels of terbutaline achieved by this technique are around 150 nmol/litre, the normal therapeutic range for oral terbutaline treatment (7.5 mg twice daily) being 10-20 nmol/litre but despite this, significant changes in serum potassium or glucose concentrations are rare as, surprisingly, is tremor which may suggest the development of tolerance to the side effects of this form of treatment. Muscle cramps are common and may sometimes be severe, with elevation of plasma creatinine phosphokinase (Sykes et al, 1991) although levels of the myocardial fraction are normal. Some patients complain of an effect on memory and ability to concentrate and occasionally menorrhagia is seen but this is not usually severe.

The main problem is the development of subcutaneous inflammatory nodules. When biopsied these show an eosinophilic infiltrate (Lewis et al, 1987). These usually settle down once that area of skin is avoided, but often leave a fibrotic nodule. More recently a more aggressive type of lesion has been demonstrated which sometimes leads on to frank abscess formation, the pus from which is usually sterile. The formulation of the drug has not changed nor have the preservatives, so the reason for these reactions remains elusive. Although using nebulizer solution rather than the injectable form of terbutaline may help, in some the skin changes are so severe that administration has to be changed to continuous intravenous infusion via an indwelling line such as a Portacath or Hickman line, although in those patients who have had many hospital admissions, such vascular access may also be needed because of lack of useable veins.

Long-acting inhaled §2 agonists
In our experience, salmeterol has proved to be disappointing in these patients for reasons that are not clear. Whether formoterol, which is a full agonist may be more useful than salmeterol, a partial agonist, remains to be determined.

Management: Type 2 brittle asthma

Management of type 2 patients is less difficult. In view of the rapid onset of attacks each patient should be provided with a medic alert bracelet or equivalent. Again identification of inhaled or ingested triggers, e.g. peanuts (Loza and Brostoff, 1995), is crucial, but the mainstay of self treatment for these attacks is adrenaline. Although these patients often appear to be relatively symptom free between attacks in some cases significant peak flow variability is seen which is not matched by perceived symptoms, which may explain the appearance of sudden attacks occurring on the background of significant but undetected airway narrowing.

Adrenaline may have theoretical advantages over selective 132 agonists, because of its action as an alpha adrenoceptor against reducing airway oedema as discussed in the section on acute airway narrowing above. Preloaded syringes (Epi-Pen, ALK, UK; Ana Pen, Allerayde, UK) should be provided for emergency treatment. Inhaled adrenaline may be more effective than a selective 132 agonist inhaler. Once adrenaline has been injected the patient should be encouraged to use a dose of nebulized salbutamol or terbutaline and go to casualty. One problem with these patients is that, once they arrive in casualty, their symptoms have often improved such that they are told either that they do not have asthma or are just hyperventilating’. Rapid onset attacks such as these are often equally quick to resolve leading to the opportunity for inadequate assessment of severity in these cases.


Patients with brittle asthma, whether type 1 or type 2, pose difficult and complex management problems. Trying to classify these severe patients will help to determine the differing factors involved and, while this classification will not embrace all patients with severe asthma, it will provide a framework for beginning to unravel aetiology and treatment of this high morbidity group. Once identified, dealing with individual factors may in themselves have only a small impact on their condition but these can be cumulative, and even if these result in only modest improvements in control, the patient will believe that some improvement can after all be achieved.


Ayres JG (1992)
Subcutaneous terbutaline in the managemeni of brittle asthma. Br J Hosp Med 47: 569-71

Baker JS, Tunnicliffe WS, Duncanson RC, Ayres JG (1995)
Reduced dietary intakes of magnesium, selenium and vitamins A, C and F in patients with brittle asthma. Thorax 50: A75

Baker JC, Tunnicliffe WS, Duncanson RC, Ayres JG (1996)
Double blind placebo controlled food challenge in type I and type 2 brittle asthma. Thorax 51: A2

Barnes PJ, Adcock IM (1995)
Steroid resistant asthma. Q J Med 88: 455-68

Bateman JRM, Clarke SW (1979)
Sudden death in asthma. Thorax 34: 40-3

British Thoracic Society, British Paediatric Association, Royal College of Physicians of London et al (1996)
Guidelines on the management of asthma. Thorax 52 (Suppl): S1-24

Garden GMF, Ayres JG (1993)
The psychiatric and social aspects of brittle asthma. Thorax 48: 501-5

Lewis LD, O’Driscoll BRC, Hartley RB, Cochrane GM (1987)
An unusual local reaction to continuous subcutaneous infused terbutaline in unstable asthmatics. Br J Dis Chest 81: 189-93

Loza C, Brostoff J (1995)
Peanut allergy. Clin Exp Allergy 25: 493-502

Miles JF, Noble K, Matthews BR, Cayson RM, Ayres JG (1993)
Gastro-oesophageal reflux in patients with brittle asthma. Thorax 48: 1055

Miles JF, Cayton RM, Tunnicliffe WS, Ayres JG (1995)
Increased atopic sensitization in brittle asthma. Clin Exp Allergy 25: 1074-82

Miles IF, Garden GM, Tunnicliffe W, Cayton RM, Ayres JG (1997)
Psychological morbidity and coping skills in patients with brittle and non-brittle asthma: a case-control study. Clin Exp Allergy (in press)

Miller MR, Dickinson SA, Hitchings DJ (1992)
The accuracy of portable peak flow meters. Thorax 47: 904-9

O’Driscoll BRC, Ruffles SP, Ayres JG, Cochrane GM (1988)
Long-term treatment of severe asthma with subcutaneous terbutaline. Br J Dis Chest 82: 360-5

Sykes AP, Lawson N, Finnegan JA, Ayres JG (1991)
Creatinine kinase activity in patients with brittle asthma treated with long term subcutaneous terbutaline. Thorax 46: 580-3

Turner-Warwick M (1977)
On observing patterns of airflow obstruction in chronic asthma. Br J Dis Chest 71: 73-86

Wasserfallen JB, Schaller MD, Feihl F, Perret CH (1990)
Sudden asphyxic asthma: a distinct entity? Am Rev Respir Dis 142: 108-11

Westerman DE. Benatar SR. Porigieter PD. Ferguson AD (1979)
Identification of high risk asthmatic patients. Am J Med 66: 565-72

Key Points

  • Brittle asthma is relatively uncommon and consequently has been characterized in clinical terms.
  • Two types at least can be identified, type 1 characterized by wide peak flow variation despite maximal therapy and type 2 by very sudden attacks.
  • Type 1 brittle asthma is associated with the female sex, atopy, high psychosocial disturbance and food intolerance. Treatment has to be holistic.
  • Type 2 brittle asthma is best treated with self-injectable adrenaline and avoidance of recognized triggers.