• 
• 

Three Part Question

In [a 2 years old child with breath holding spells] will a [treatment with iron] [reduce the frequency of episodes]?

Clinical Scenario

A 2 year old child is seen in the out patients department with a history of breath holding spells for the last three months, occurring about 3-4 times per week. These are causing her mother a great deal of concern. You consider whether or not a course of iron would reduce the frequency of these attacks.

Search Strategy

Medline 1996-2001, using the OVID interface
Breath holding spells AND iron therapy LIMIT to English AND Child

Search Outcome

6 hits: Two letters, an editorial and three papers as discussed below. Secondary Searches -Cochrane, Clinical Evidence -none.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Bhatia MS et al, 1990, India	50 children with BHS	Case Series (4)	A reduction in frequency and severity of BHS following a course of iron which was continued until Hb>11g/dl	96% of the study group had IDA (mean Hb=8.12g/dl). 82% showed a response within 2 weeks. After three weeks 100% of cases had shown an improvement	Failed to define improvement
Daoud AS et al, Jordan, 1997	67 children with BHS were randomised to either ferrous sulfate or placebo	RCT (level 1b)	Frequency of BHS per month after 16 weeks. Complete response (no attacks), partial response (>/= 50% reduction), minimal or no response (<50% reduction)	Complete response in 51.5% treated vs. 0% in non-treatment group. Partial response in 36.4% treated group vs. 5.9% in non-treatment group. NNT for any response = 2 (95% CI 2 to 3)	The baseline mean Hb in children who showed a response was 86g/l compared to 106g/l in those that responded poorly (p=0.004)
Mocan H et al, Turkey, 1999	91 children with breath holding spells. 63 with concomitant IDA	Case Control (4)	Frequency of BHS over the 3 months study period. Complete response (no attacks), partial response (>/=50% reduction), no response	Complete response in 32% treated group and 21% more had partial response. NNT (for either a partial or complete response) = 2 (95% CI 2 to 3)	Only those children with BHS and IDA were treated with iron The placebo group were those with BHS but normal Hb

Comment(s)

The literature reviewed suggests that a trial of iron therapy will reduce the frequency of breath holding spells. All these papers showed a high incidence of iron deficiency anaemia associated with breath holding spells. A full blood count would therefore be warranted in the work up of these children. Treatment is more likely to be successful when there is concomitant iron deficiency anaemia. Length of treatment varied between 3 and 16 weeks with ferrous sulphate (5-6mg/kg/day). A course of 8 weeks would seem reasonable - long enough to improve any anaemia. There was no mention of side effects with ferrous sulphate treatment in any of these papers. Typically these would include nausea, vomiting, diarrhoea and change in stool colour; the latter presumably making it difficult to complete a double blind study of iron therapy. There is also the risk of accidental overdose by the patient or siblings to be considered. However the risk of overdose with paediatric preparations of iron is minimal compared with adult preparations. The decision to treat also needs to be balanced against the natural course of breath holding spells which are invariably benign and self limiting - both in the short and long term

Clinical Bottom Line

Iron therapy is of benefit in children with breath holding spells (NNT=2). Improvement is more likely in those with concomitant iron deficiency anaemia (NNT=1).

References

1. Bhatia MS, Singhal PK, Dhar NK, et al. Breath holding spells: an analysis of 50 cases. Indian Pediatr 1990;27(10):1073-9.
2. Daoud AS, Batieha A, al-Sheyyab M, et al. Effectiveness of iron therapy on breath-holding spells. J Pediatr 1997;130(4):547-50.
3. Mocan H, Yildiran A, Orhan F, et al. Breath holding spells in 91 children and response to treatment with iron. Arch Dis Child 1999;81(3):261-2.