Off-label drug use and the risk of medication errors in critically ill neonates: A conceptual pilot study

Mean (SD) of serum creatinine amongst the study population was 64.7 (16.8) μmol∕L and the corresponding eGFR was 35.1 (12.8) ml∕min∕1.73 m². All components for assessing the MERIS score were available for 148 participants. The MERIS evaluation revealed average (SD) scores for the number of drugs, eGFR and harm/interaction of 15.6 (4.8), 17.2 (3) and 11.2 (2.7) respectively with a final score of 41 (7.7). Assessment of the total and individual components of MERIS according to the gestational age and birth weight categories are depicted in Fig. 4. Significantly higher scores for all the sub-components as well as for the total MERIS was observed for younger neonates as well as lower birth weights. Out of the 148 neonates evaluated for MERIS, 45 received only labelled drugs while the remaining at least one OLD/unlicensed drug. Significantly (p = 0.01) higher scores except in eGFR category were observed for neonates receiving at least one of either off-label drug or unlicensed drug compared to those receiving only labelled drugs (Fig. 5). Out of 103 neonates with at least one OLD/unlicensed drug use, 84 (81.6%) had higher risks of medication error as per MERIS compared to only 8/45 (17.8%) with labelled drugs; and this was statistically significant (p < 0.0001). Similarly, higher proportions of very pre-term and extreme pre-term neonates (also very low and extremely low birth weight) neonates had significantly (p = 0.0001) more risk of medication errors (with MERIS score >13) compared to term and late pre-term neonates (Fig. 6). Binary logistic regression analysis revealed a trend between the use of OLD/unlicensed drugs and the risk of medication errors compared to use of only labelled drugs [OR:20.4, 95%CI: 8.2, 50.9; p = 0.0001].

OPEN IN VIEWER

Fig. 4.

Medicine risk score (MERIS) in each sub category of the study population. Statistically significant (= 0.0001) differences were observed in all scores between various categories of gestational age and birth weight. PT-Pre-term; NBW-Normal birth weight; LBW-Low birth weight; VLBW-Very low birth weight; ELBW-Extremely low birth weight.

OPEN IN VIEWER

Fig. 5.

Comparisons of the individual components and total medication risk score (MERIS) between neonates with at least one off-label drug compared to those with only labelled drugs. Significant differences (p < 0.05) were observed in all scores except for the eGFR score between neonates receiving only labelled compared to those with at least one of the OLDU or unlicensed drugs. OLDU-Off-label drug use; MERIS-Medication risk score.

OPEN IN VIEWER

Fig. 6.

Comparison of risk of medication errors in various gestational age groups and birth weight categories. Significant proportions (p < 0.05) of very and extremely, and very low and extremely low birthweight neonates had medication risk scores (MERIS) ≥13 compared to other groups.

The American Academy of Pediatrics in its policy statement has stated that “off-label drug use does not constitute improper, illegal, contraindicated or investigational use” [13]. USFDA approves a drug and drug label based on substantial evidence based on adequate and well-controlled evidence [23]. In children, OLD use has been observed to triple the risk of adverse drug reactions [24]. We observed that majority of the OLD/unlicensed drug use were drugs belonging to dermatology and nervous system categories according to WHO ATC classification. This is similar like studies from European and Australian intensive care units [25–27]. However, drugs in the category of alimentary tract and metabolism was reported with the highest frequency of being used off-label from USA; and anti-infective drugs from developing countries [4,28]. Pharmaceutical companies have stepped forward in conducting more paediatric studies as evident by the increased clinical trial reports, but it is certainly a long way to go in neonates [29]. Several regulations such as European Union Pediatric regulation have hardly made any significant impact in the prevalence of OLD use in children [30].

Differences in the drug regulatory agencies in drug approvals and consequent label information were also observed in previous studies [31]. American Academy of Paediatrics insists on obtaining consent from the parents if a drug is used either in an off-label manner or if it is unlicensed [32]. Partly, onus lies with the practising physicians in publishing their experience with such drugs that adds to the scientific literature and may aid in therapeutic decision-making. Pediatric advocacy groups must continue insisting label changes in neonatal population.

Medication errors have been reported to occur at an incidence between 5–13% in NICU [33–35]. The Institute for Safe Medication Practice has classified 10 drugs/drug classes as high alert medications as they are highly susceptible to dosing errors with serious consequences [36]. Using the MERIS tool, we observed that nearly 61.5% of neonates in the present study had potential risk of medication errors. Using the MERIS tool, we observed that the odds of medication error risk may be 20.4 times higher than using only labelled drugs.

While applying the MERIS tool in the neonatal population, we observed a limitation that one of the components, eGFR score, was not discriminatory between the sub-populations and did not contribute significantly to the total scores. In the present study, eGFR was based on serum creatinine. We used the Shull’s equation in the present study for evaluating eGFR as we did not have length measurements of the neonates. Assessing the eGFR with one of the equations containing length of the neonates may improve the utility of this item in the scoring tool. Several novel biomarkers such as cystatin C can accurately predict the glomerular perfusion rate in neonates, it is often not practically feasible in developing countries to inculcate this as the standard of care and hence serum creatinine based equations are adhere to as in the present study. However, neonatal specific eGFR based on a validated biomarker would add more value in developing such scoring systems. Alternately, number of drugs with potential renal toxicity may be considered instead of eGFR score. Although MERIS has been developed and validated in adult population, we do not have any validated paediatric/neonate specific medication error tool. Apart from this, the study is also limited in not exploring the level of scientific evidence for OLD use and the mentions in drug compendia such as Micromedex or British National Formulary; and no validated scoring system was used to assess the severity of illnesses. Also, the classification list of drugs for assessing the MERIS tool based on their potential to cause harms/interaction was validated by four neonatologist and one clinical pharmacologist. Amongst the neonatologist, the experience in the field ranged between 10–26 years and the clinical pharmacologist had 12 years of experience. It is preferable to validate the tool with large network of potentially diverse population involved in neonatal care. In the original article related to the MERIS tool, during the initial phase of model development, co-morbidity was considered as one of the elements that was subsequently left. Future studies validating such a tool in the neonatal population should consider co-morbid diseases as an element.

5. Conclusion

In conclusion, significant proportions of critically ill neonates received either off-label or unlicensed drug and consequently were associated with a potential risk of medication errors. However, it must be noted that we did not actually measure the medication errors but the potential risk. We observed that the eGFR component based on serum creatinine of the MERIS score that was adapted in the presents study, does not discriminate between those with the risk of medication errors and those without. We recommend validating a modified MERIS tool using alternate measuring method for eGFR based on novel biomarkers or using scoring systems that include length as one of the variables as stated above. This is critical in neonates considering the rapid maturation of nephrons during the post-natal period. Developing neonate-specific tool for assessing the risk of medication errors may aid in better understanding the characteristics of neonates and drugs/drug classes that are more prone for medication errors. The results of this study should be considered as preliminary evidence in identifying that there is a potential risk of medication errors particularly with OLDU in neonates.