• 
• 

Three Part Question

In [children with spontaneous epistaxis] which [clinical findings] predict [coagulation or platelet abnormalities]?

Clinical Scenario

A four year old girl is brought to the Paediatric Emergency Department by her family with her second episode of spontaneous epistaxis in a month. The bleeding resolves within 30 minutes and she is otherwise fit and well. Her parents are requesting blood tests to look for an underlying cause but you are uncertain as to whether this is justified.

Search Strategy

Embase (1980 to present) and MEDLINE (1950 to present) databases searched via NHS Evidence Health Information Resources on 2nd May 2012
Search terms used:
[child* OR paediatric* OR pediatric*] AND [epistaxis* OR nosebleed* OR nasal AND bleeding*] AND [coag* OR platelet* OR clotting*].

Search Outcome

210 papers were identified of which 3 were relevant and of sufficient quality for inclusion.
The references of identified articles were searched manually and 4 further papers were found suitable for inclusion.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Katsanis et al 1988 Canada	36 children with 5 or more spontaneous nosebleeds (1986-1987) referred from ED or patient’s physician	Prospective observational study	Development of an epistaxis scoring system to assess severity of recurrent nosebleeds and identify children needing full haemostatic workup.	Abnormal coagulation tests correlated with epistaxis scored as severe. (p<0.02 for prolonged bleeding time and p<0.05 for VWFAg and VWFRisocetin) Positive bleeding history (excluding epistaxis) (p<0.05), family history (p<0.01) and need for cauterisation (p<0.01) were found to be statistically significant in severe epistaxis	Historical data. No details of derivation of scoring system. No explanation of use of additional features such as bleeding history and family history in score Elements of epistaxis scoring system difficult in clinical practice such as estimation of blood loss. Variability of VWF levels in patients due to role as acute phase protein makes diagnosis or exclusion difficult on single test. Recommendation of extended coagulation screen may not be appropriate in emergency department without haematological advice.
Sandoval et al 2002 USA	178 children referred for the evaluation of recurrent epistaxis to outpatient haematology clinic (1985-1999)	Retrospective observational study	To determine the clinical and laboratory features of children with recurrent epistaxis	59 children (33%) had a coagulopathy. Family history was predictive of diagnosing a coagulopathy (P=0.023) Duration and severity of epistaxis and presence of other bleeding symptoms had no predictive value.	Ascertainment bias (Haematology Clinic). Single centre. Retrospective
Brown et al 2004 Australia	14 patients admitted with acute epistaxis in tertiary pediatric centre (1992-2002)	Retrospective observational study	To determine the outcomes for healthy children who require admission to hospital with acute epistaxis	No patient diagnosed with bleeding disorder.	Ascertainment bias (admitted patients). Small numbers. Differing laboratory reporting methodology over study period.
Damrose et al 2006 USA	90 children referred for outpatient evaluation and treatment of epistaxis (January 2000 to January 2003)	Retrospective observational study	To evaluate approach used to evaluate and treat patients referred with epistaxis on an outpatient basis	Abnormal coagulation values were identified in 7.8% (n=7) of patients. Of these: • 4 were diagnosed with specific coagulopathies. • 3 had a positive family history of bleeding disorder. Most cases resolved with emollient	Ascertainment bias (ENT outpatients). Single centre, single physician study. No estimate of predictive value of clinical features in detecting coagulopathy. Limited coagulation tests performed. No comparison of clinical features in children with or without coagulopathies.
Bowman et al 2009 Canada	151 children under investigation for Von Willebrand Disease (VWD)	Prospective observational study	Prospective validation of Pediatric Bleeding Questionnaire.	Sensitivity, specificity, positive predictive value and negative predictive value were: 83%, 79%, 0.14 and 0.99 respectively. ROC analysis showed questionnaire can distinguish between affected and unaffected children. Specific to epistaxis, longer duration (>10 minutes) , lack of seasonal correlation and need for medical intervention were associated with VWD (p<0.05)	Only considered diagnosis of VWD and did not investigate for other disorders.. Full paediatric bleeding questionnaire impractical for use in emergency department as can take up to 40 minutes to complete. Very wide 95% CIs for sensitivity. Ascertainment bias – investigated because personal history of haemorrhagic symptoms and/or family history of VWD and/or for pre-operative screening
Paranjothy et al 2009 Wales	36 infants admitted to hospital with epistaxis (January 1999 to December 2004)	Retrospective observational study	To estimate the incidence and describe the aetiology of epistaxis in infants	4 patients (11.1%) had coagulation disorder.	Ascertainment bias (Admitted patients). Only looked at epistaxis in under ones. Not all patients had coagulation or platelet count checked. No consideration of predictive features.
Elden et al 2011 USA	47 children referred to otolaryngology (October 2006-December 2010)	Retrospective observational study	To identify the prevalence of previously undiagnosed bleeding disorders in children with severe epistaxis requiring intraoperative nasal cautery	15 children had abnormal coagulation studies, 12 were referred to haematology, of which 5 (10.6% of total study group) diagnosed with bleeding disorder	Ascertainment bias (ENT outpatients who failed medical therapy). Single centre. Small numbers. Clinical features not identified Abstract only

Comment(s)

Epistaxis is a relatively common occurrence in childhood often due to local trauma (accidental or nose picking) but can also be associated with upper respiratory tract infections and some underlying systemic causes including bleeding disorders. The majority of nosebleeds are spontaneous, short, and self-limiting, but parental concern may be heightened when they are of longer duration and recurrent. The clinical conundrum lies in determining which patients need investigation. The studies above show that up to one third of children with spontaneous epistaxis may have an underlying bleeding disorder. However, these were performed in specific subgroups of patients (admitted to hospital, referred to ENT/haematology clinics) and are thus unlikely to be representative of patients presenting directly to an Emergency Department (ED). The only consistently identified predictor of coagulopathy was a positive family history. Other predictors proposed were longer duration (>10 minutes) and personal history of bleeding (excluding epistaxis) although these were not replicated throughout the studies. Lack of seasonal correlation and need for medical intervention may also be predictors. Whilst there has been derivation and validation of paediatric bleeding scores, these are neither specific to epistaxis nor practical for use in an ED given their complexity. They do however highlight the need to explore other bleeding episodes in the history. One paper found that most cases resolve with topical emollients. Failure to respond to this may prove to be a discriminator in identifying those who warrant investigation though no further evidence exists to support this as yet. The emphasis must therefore remain on full medical assessment and consideration of underlying disorders. We have not assessed the literature for links to any other underlying disorder, including hypertension. Where there is concern regarding possible bleeding disorders prompt referral must be initiated and investigations undertaken with specialist advice as necessary.

Clinical Bottom Line

There is insufficient evidence to determine which clinical features necessitate further investigation in children with spontaneous epistaxis. The most useful predictor appears to be a positive family history, with other criteria being less useful. Failure of emollient therapy may be helpful in determining which children require onward referral.

Level of Evidence

Level 3 - Small numbers of small studies or great heterogeneity or very different population.

References

1. Katsanis E, Koon-Hung L, Hsu E et al. Prevalence and Significance of Mild Bleeding Disorders in Children with Recurrent Epistaxis. The Journal of Pediatrics. 1988; 113(1):73-76.
2. Sandoval C, Dong S, Visintainer P et al. Clinical and Laboratory Features of 178 Children With recurrent Epistaxis. Journal of Pediatric Hematology/Oncology. 2002; 24(1): 47-49.
3. Brown NJ, Berkowitz RG. Epistaxis in healthy children requiring hospital admission. International Journal of Pediatric Otorhinolaryngology. 2004; 68;1181-1184.
4. Damrose JF, Maddalozzo. Epistaxis. Laryngoscope 2006; 116:387-393.
5. Bowman et al. Evaluation of the Diagnostic Utility for von Willebrand Disease of a Pediatric Bleeding Questionnaire Journal of Thrombosis and Haemostasis. 2009; 7:1418-21.
6. Paranjothy S, Fone D, Mann M, Danstan F et al. The incidence and aetiology of epistaxis in infants: a population-based study. Arch Dis Child 2009; 94:5421-424.
7. Elden LM, Reindeers M, Witmer C. Predictors of Bleeding Disorders in Children with Epistaxis: Value of Preoperative Tests and Clinical Screening. Otolaryngology – Head and Neck Surgery 2011; 145:243.