Off label prescribing to children in primary care in Germany: retrospective cohort study

Between 35% and 90% of the drugs prescribed to hospitalised children are either not licensed for children's use or are prescribed outside the terms of their product licence (off label prescribing).¹,² Subsequent adverse reactions are more likely than with licensed products (6.0% v 3.9%).³ We analysed the extent of prescribing off labelled products in a representative cohort of children in primary care.

Patients, methods, and results

We used the electronic database of prescriptions of Allgemeine Ortskrankenkasse, Baden-Württemberg. This health insurer covers more than four million people, 42% of the total population of the state. We retrospectively reviewed 1.74 million anonymous prescriptions written by 6886 office based doctors—specialists in paediatric, general, or internal medicine—between 1 January and 31 March 1999 for 455 661 patients aged 0-16 years.

Each prescription was represented by a numerical code, describing the drug's brand name, generic name, formulation, and content per dose unit. Our database did not contain diagnoses, dosage recommendations, or individually prepared drug formulations.

To assess the licence status of prescriptions we used the summary of product characteristics (Fach information) or drug lists provided by German pharmaceutical manufacturers' associations (Gelbe Liste or Rote Liste).

We categorised prescriptions by age group and the World Health Organization's anatomical, therapeutic, and chemical classification. A prescription was considered off label if the drug itself, its dose unit, or its formulation was not explicitly covered by documentation for the specific age group to which it was prescribed. Unlicensed drugs are not specified in the database because they are not automatically reimbursed by insurance.

Of 1740 238 prescriptions, 115 366 (6.6%) prescriptions for medical accessories, diets, and cosmetics and 32 866 with unidentifiable codes were excluded; the prescriptions with unidentifiable codes might have included an unknown number of unlicensed prescriptions but accounted for only 1.9% of the database.

Among the remaining 1 592 006 prescriptions for 10 452 different active ingredients, we found 210 528 (13.2%, 95% confidence interval 13.2% to 13.3%) off label prescriptions. The table shows the most common examples and some of the associated risks.

Three quarters of off label prescriptions (157 951) resulted from lack of information about use of the drugs among children or in particular age ranges. Of the off label prescriptions, 35 234 (16.7%) ignored recommendations on active ingredient, dose units, or formulations for a specific age group—for example, quinolones in children and xylometazoline 1% formulations for babies.

The proportion of off label prescriptions was highest for 1-2 year olds (68 791 (17.9%, 17.8% to 18.1%) prescriptions) and lowest for 7-11 year olds (40 539 (10.5%, 10.4% to 10.6%) prescriptions).

Of the 181 914 (8.8%) prescriptions for topical treatments of the skin, eye, or ear, 116 060 (63.8%, 63.6% to 64.0%) were off label. The active ingredients of the most commonly prescribed systemic off label drugs are shown in the table.

Off label prescribing was common for cardiovascular drugs (3646; 55.2%, 53.9% to 56.4%), drugs for genitourinary disorders (1869; 48.5%, 46.9% to 50.1%), anti-inflammatory agents (7194; 45.0%, 45.2% to 46.0%), antidepressants (246; 36.6%, 33.0% to 40.4%), and antidementia (11; 34.4%, 18.6% to 53.2%), antiepileptic (932; 14.2%, 13.3% to 15.0%), and antipsychotic drugs (54; 10.2%, 7.8% to 13.2%).

Comment

We found that 13.2% of prescriptions for a representative group of children in primary care in Germany were off label. Although we could not detect off label use due to dosage or indication with this database, the proportion of prescriptions that were off label was similar to that in much smaller studies that analysed dosage and diagnoses.⁴,⁵ Our data show that efforts to improve the quality of pharmacotherapy in children should not exclude widely marketed and firmly established drugs.

Table.

Most frequent off label drugs prescribed to outpatients aged 0-16 years for peroral, rectal, or nasal administration at the expense of Allgemeine Ortskrankenkasse, Baden-Württemberg, between January and March 1999

Rank	0-11 months			1-2 years			3-6 years			7-11 years			12-16 years
	Drug	No		Drug	No		Drug	No		Drug	No		Drug	No
1	Xylometazoline or oxymetazoline*	217		Xylometazoline or oxymetazoline*	13 780		Xylometazoline or oxymetazoline*	3524		Cetylpyridinium‡‡	2651		Tyrothricin mixtures‡‡	4234
2	Herbal extract of ivy†	149		Saccharomyces boulardii	3 611		Dihydrocodeine**	2921		Sultiame	440		Cetylpyridinium mixtures‡‡	3132
3	Pipenzolate	145		Salbutamol§	2 394		Tyrothricin mixtures‡‡	1470		Codeine mixtures**	268		Ibuprofen*	1942
4	Saccharomyces boulardii	36		Mucolytic herbal formulations†	1 018		Cetylpyridinium‡‡	664		Formoterol	243		Diclofenac‡‡	1128
5	Acetylcysteine	32		Codeine mixtures**	730		Loratadine	477		Mucolytic herbal formulations†	232		Magaldrate	327
6	Cisapride‡	30		Dihydrocodeine**	687		Pipenzolate	415		Diclofenac‡‡	227		Sultiame	220
7	Salbutamol§	24		Doxylamine mixtures (with or without paracetamol)	672		Fluticasone propionate	404		Echinacea purpurea formulations†	211		Extract of Lichen islandicus†	214
8	Terbutaline§	19		Pipenzolate	486		Mucolytic herbal formulations†	276		Extract of Lichen islandicus†	200		Mucolytic herbal formulations†	205
9	Antacids	19		Tetryzoline*	271		Diclofenac‡‡	238		Dihydroergotamine	169		Crataegus and camphor formulations†	205
10	Ofloxacin¶	13		Cisapride‡	229		Ofloxacin¶	155		Antacids	156		Ofloxacin¶	179

^*Prescribed amount of drug per dose exceeded the recommended dose. 

^†No dosage recommendations were available. Herbal formulations containing as much as 65% of ethanol by volume may cause significant concentrations of ethanol in babies and small children. 

^‡Cisapride is known to induce cardiac arrhythmias. It has been withdrawn. 

^§Efficacy and safety of β-2-sympathomimetics have not been proved in children younger than 18 months. 

^¶Use of the quinolone ofloxacin is not recommended during growth. 

^**For this age group no dosage is indicated in the SPC due to lacking pharmacokinetic data. Doses >3 mg/kg/day have been observed to produce respiratory depression, somnolence, or vomiting. 

^‡‡Due to a lack of data, there are no dosage recommendations for children younger than 15 years, when diclofenac is administered systemically. 

Editorial by Banner and pp 1312, 1313