Changes in Dispensing of Acid Blockers to US Children: 2016–2023

Acid blockers, defined as H2 receptor antagonists (H2RAs) and proton pump inhibitors (PPIs), are among the most commonly used medications in children.¹ These drugs are the primary pharmacologic treatments for gastroesophageal reflux disease, a condition affecting approximately 2% to 8% of children.^2,3 Recent events may have affected acid blocker dispensing to US children. First, in late September 2019, several manufacturers voluntarily recalled the most commonly used H2RA, ranitidine, owing to contamination with a carcinogen.⁴ On April 1, 2020, ranitidine was withdrawn from the market.⁵ Second, after the COVID-19 pandemic began in March 2020, pediatric outpatient visits declined sharply,⁶ potentially decreasing opportunities to initiate acid blocker therapy.

We are unaware of any studies that have evaluated changes in acid blocker dispensing to children after these 2 events. Evaluating these changes could provide insights on the effects of drug recalls and the COVID-19 pandemic on pediatric prescribing patterns, topics on which there are little data. From the perspective of health care quality, it is important to evaluate changes in acid blocker dispensing to infants specifically. While these medications can prevent complications from severe gastroesophageal reflux disease such as failure to thrive,² evidence suggests that many infants prescribed acid blockers—if not most—have uncomplicated physiologic gastroesophageal reflux.^7–10 For such infants, lifestyle interventions, such as maternal elimination diets or thickened feedings, should be implemented instead of prescribing acid blockers, which do not improve symptoms and may increase the long-term risk of infection and allergic diseases.^11–13 

In this study, we used a quasi-experimental design and 2016–2023 data from a comprehensive national prescription dispensing database to evaluate the association between the ranitidine recall and the beginning of the COVID-19 pandemic on acid blocker dispensing to US children. Additionally, we conducted subgroup analyses among infants and older children.

We analyzed the 2016–2023 IQVIA Longitudinal Prescription Database. This all-payer database reports prescription dispensing from 92% of US retail pharmacies and the majority of mail-order and long-term care pharmacies. Data elements used in this analysis include an encrypted patient identifier, year of birth, and national drug code. The database does not report patient race and ethnicity, income, or prescription indication. Because data were deidentified, the Institutional Review Board of the University of Michigan exempted this study from review. This study follows the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for reporting cross-sectional studies.¹⁴ 

The sample included acid blocker prescriptions dispensed to children aged 0 to 18 years between January 1, 2016, and December 31, 2023. We used IQVIA’s market definition to identify these prescriptions. H2RAs primarily included ranitidine and famotidine; cimetidine and nizatidine were included for completeness but were rarely dispensed during the study period. PPIs included omeprazole, lansoprazole, pantoprazole, esomeprazole, and rabeprazole.

Outcomes were the monthly number of dispensed prescriptions for any acid blocker, any H2RA, and any PPI. We fitted segmented regression models assessing for level and slope changes in outcomes during October 2019 (the month after the ranitidine recall) and April 2020 (the month after the COVID-19 pandemic began).¹⁵ To account for seasonality, models controlled for quarter. To facilitate interpretation, we calculated the difference between the outcome in December 2023 and the outcome predicted by the counterfactual trend (ie, the trend that would have occurred had trends from January 2016-September 2019 continued). We accounted for autocorrelation using robust Newey-West standard errors with the appropriate number of lags.¹⁶ 

To identify infants and older children for subgroup analyses, we ideally would have had access to information on date of birth, allowing us to define infants as those who were aged less than 1 year at the time of the fill and older children as those who were aged 1 year or older, following the age dichotomization used in a national guideline for the management for gastroesophageal reflux.¹³ However, because we only had information on year of birth, we could only calculate age in years as of December 31. To ensure capture of all acid blocker dispensing to patients who were aged less than 1 year at the time of the fill, we chose to define infants as those aged 0 to 1 year as of December 31 and older children as those aged 2 to 11 years as of December 31. A drawback of this analytic choice was that some prescriptions assigned to the infant group were for patients who were aged 1 year or older at the time of the fill.

To evaluate whether secular trends in prescription drug dispensing to children may have affected results, we repeated the interrupted time series analysis when evaluating the monthly number of dispensed prescriptions for drugs other than acid blockers. We also evaluated whether there were changes in monthly mean out-of-pocket spending per dispensed acid blocker prescription during the study period, as cost sharing may impede prescription drug dispensing. Analyses used R version 4.2.2 and 2-sided hypothesis tests with α = 0.05.

Figure 1 displays monthly dispensing of ranitidine, famotidine, other H2RAs, and PPIs to children during 2016 to 2023. Ranitidine dispensing declined from 215 229 to 210 221 prescriptions between January 2016 and September 2019, decreased rapidly from October 2019 through April 2020, and virtually ceased afterwards.

Monthly famotidine dispensing ranged between 23 465 and 31 206 prescriptions between January 2016 and September 2019. After this month, famotidine dispensing increased sharply, reached a peak of 217 489 prescriptions in March 2023, and declined to 194 039 prescriptions in December 2023. Dispensing of other H2RAs (cimetidine and nizatidine) was infrequent throughout the study period.

PPI dispensing declined from 227 836 to 177 396 prescriptions between January 2016 and September 2019. In October 2019, PPI dispensing transiently increased to 200 805 prescriptions and then declined to 150 162 prescriptions in December 2023.

In January 2016, the predicted number of dispensed prescriptions for acid blockers to children (intercept) was 505 161 (95% CI, 490 081 to 520 240). Before October 2019, this number declined by 1330 prescriptions per month (95% CI, −1842 to −818). In October 2019, there was a large level decrease of 55 809 prescriptions (95% CI, −92 038 to −19 580), or a 11.0% decline relative to the intercept. The slope change after October 2019 was negative but not significant (−9196 prescriptions per month; 95% CI, −20 350 to 1958). In April 2020, there was no significant level change, but there was a slope increase of 11 749 prescription per month (95% CI, 595 to 22 903). Despite this increase, the number of dispensed prescriptions for acid blockers to children in December 2023 was 42 219 (−10.9%) lower than predicted had pre-October 2019 trends continued (Table 1; Figure 2a).

In January 2016, the predicted number of dispensed prescriptions for H2RAs to children was 268 674 (95% CI, 259 354 to 277 995). Before October 2019, this number changed little from month to month. In October 2019, there was a large-level decrease of 70 584 prescriptions (95% CI, −92 997 to −48 191), representing a 26.3% decline relative to the intercept, as well as a slope decrease of −11 017 prescriptions per month (95% CI, −17 912 to −4123). In April 2020, there was no level change, but there was a slope increase of 12 673 prescriptions per month (95% CI, 5779 to 19 568). Despite this increase, the number of dispensed prescriptions for H2RAs to children in December 2023 was 81 886 (−29.5%) lower than predicted had pre-October 2019 trends continued (Table 1; Figure 2b).

In January 2016, the predicted number of dispensed prescriptions for PPIs to children was 236 486 (95% CI, 229 981 to 242,991). Before October 2019, this number declined by 1422 prescriptions per month (95% CI, −1643 to −1201). In October 2019, there was a nonsignificant level increase of 14 775 prescriptions (95% CI, −853 to 30 403) but no slope change. There was no level or slope change in April 2020. In December 2023, the number of dispensed prescriptions for PPIs to children was 39 667 (35.9%) higher than predicted had pre-October 2019 trends continued (Table 1; Figure 2c).

Figure 3a shows changes in acid blocker dispensing to infants and older children in October 2019 and April 2020. For infants, there was a −36 449 (95% CI, −46 643 to −26 355) prescription level decrease in acid blocker dispensing in October 2019, or a 26.1% decline relative to the intercept. In contrast, the level change in acid blocker dispensing to older children in October 2019 was negative, although not significant (−19 310; 95% CI, −46 564 to 9945; 5.3% decline relative to the intercept). For both infants and older children, there was no significant slope change in October 2019 and no significant level or slope change in April 2020, although the point estimates for the slope changes were large and positive in both age groups. In December 2023, acid blocker dispensing to infants was 29 319 (−25.5%) prescriptions lower than predicted had pre-October 2019 trends continued. In contrast, acid blocker dispensing to older children was only 12 900 (−4.7%) prescriptions lower than predicted (Table 1).

Figure 3b shows changes in H2RA dispensing to infants and older children in October 2019 and April 2020. For both groups, there were level and slope decreases in H2RA dispensing in October 2019, no level changes in April 2020, and slope increases in April 2020. In December 2023, H2RA dispensing to infants and older children was 31.7% and 28.0% lower than predicted had pre-October 2019 trends continued, respectively.

Finally, Figure 3c shows changes in PPI dispensing to infants and older children in October 2019 and April 2020. Among infants, there was a level increase (4664; 95% CI, 3008–6281) and slope increase (527 prescriptions per month; 95% CI, 23–1031) in PPI dispensing in October 2019. In contrast, there was neither a level nor slope change in PPI dispensing to older children in October 2019. In both groups, there was no level or slope change in PPI dispensing in April 2020. In December 2023, PPI dispensing to infants and older children was 82.5% and 31.3% higher than predicted had pre-October 2019 trends continued, respectively.

Among both infants and older children, there was no level or slope change in the monthly number of dispensed prescriptions for drugs other than acid blockers in October 2019. In both groups, there was a level decrease but no slope change in April 2020. After April 2020, dispensing of drugs other than acid blockers increased in both groups (slope after April 2020 in infants: 17 082 prescriptions per month; 95% CI, 12 349–21 815; slope after April 2020 in older children: 158 020 prescriptions per month; 95% CI, 103 774–212 266). By December 2023, the number of dispensed prescriptions for drugs other than acid blockers was 16.3% and 7.9% higher than predicted by pre-October 2019 trends among infants and older children, respectively (Table 1; Figure 4). Monthly mean out-of-pocket spending per dispensed acid blocker prescription was low throughout the study period and did not change meaningfully ($8.6 in January 2016 vs $6.6 in December 2023).

Using a comprehensive, all-payer national database, we evaluated changes in acid blocker dispensing to US children after the ranitidine recall and the beginning of the COVID-19 pandemic. We found that the ranitidine recall was associated with a large and immediate decline in acid blocker dispensing. Moreover, we found that the level of acid blocker dispensing in December 2023 was still 10.9% lower than the level predicted by pre-October 2019 trends despite a recovery after April 2020. Subgroup analyses revealed that this recovery was more robust in older children than in infants, for whom acid blocker dispensing in December 2023 was 25.5% lower than the level predicted by pre-October 2019 trends.

The findings of this study suggest that drug recalls can result in large and immediate changes in dispensing patterns among pediatric patients. We believe this observation is novel, as we are unaware of any prior study evaluating changes in dispensing after a widely prescribed drug to pediatric patients was recalled and ultimately withdrawn due to safety concerns. A related body of literature has evaluated the association between black-box warnings and the prescribing of antidepressants and codeine to pediatric patients.^17,18 These studies report that black-box warnings are associated with reductions in prescribing, consistent with the notion that drug safety concerns can powerfully influence prescribing decisions to pediatric patients.

After the ranitidine recall, famotidine dispensing increased, as did PPI dispensing. These findings suggest that some clinicians and families responded to the recall by substituting other drugs. While the substitution of famotidine for ranitidine may be of little clinical consequence, the substitution of PPIs could be concerning if these drugs are associated with more adverse events compared to H2RAs. In support of this possibility, one study found that PPI therapy in young children was associated with an increased risk of fracture in young children, although this was not the case for H2RA therapy.¹⁹ 

In December 2023, dispensing of drugs other than acid blockers to infants was 16.3% higher than predicted by pre-October 2019 trends compared to 25.5% lower than predicted for acid blockers. This finding suggests that the persistent decline in acid blocker dispensing to infants was not driven by a secular trend toward decreased prescription drug dispensing to infants in general. Moreover, the fact that acid blocker dispensing to older children in December 2023 was only 4.7% lower than predicted suggests that the factors influencing the persistent decline in acid blocker dispensing to infants were specific to that age group. One possibility is that the ranitidine recall may have raised concerns about the safety of acid blockers as a class in infants, prompting some families and clinicians of infants to avoid these medications altogether. Other potential explanations include the growing awareness of the lack of efficacy of acid blockers in most infants, particularly in light of the publication of national guidelines in 2018 discouraging routine use of these drugs in this age group.¹³ Future research should investigate these possibilities.

The implications of the persistent decline in acid blocker dispensing to infants depend partly on whether it represents a decline in appropriate vs inappropriate use of acid blockers. While our database lacked the clinical details needed to determine appropriateness, studies suggest that infants are often unnecessarily prescribed acid blockers for physiologic reflux.^9,11,12 Therefore, the observed decrease in acid blocker dispensing to infants could represent a reduction in low-value care and consequently an improvement in the quality of care.

This study’s findings have clinical and policy implications. In advance of drug recalls or shifts in drug supply, professional pediatric clinical societies could convene expert working groups to provide evidence-based suggestions on patient education and clinical management, including the appropriateness and risks of medication substitutions.²⁰ For example, in response to the amoxicillin shortage in 2022, the American Academy of Pediatrics issued guidelines that highlighted safe and effective alternatives to this drug.^21,22 Additionally, by highlighting the power of drug recalls to immediately change pediatric practice, our study also suggests that securing the safety of the drug supply chain may be an important priority for pediatric health policy.

This study has limitations. First, while the study database is the most comprehensive prescription dispensing database widely available to researchers, it does not capture all prescriptions dispensed in the United States. However, we have no reason to believe that the degree of non-capture changed over time, potentially biasing results. Second, the database captures over-the-counter acid blockers prescribed by a clinician and dispensed by a pharmacist but not over-the-counter acid blockers purchased at the pharmacy register or online (eg, over-the-counter omeprazole tablets or oral dissolvable tablets). These over-the-counter formulations are likely used more often by older children than infants, suggesting that the subgroup analysis of older children may be more affected by this limitation than the subgroup analysis of infants. Third, it was beyond the scope of this study to conduct subgroup analyses among patients initiating vs continuing acid blocker therapy. Fourth, owing to the lack of information on date of birth in the database, we could only calculate age in years as of December 31. Because we chose to define infants as patients aged 0 to 1 year as of December 31, some prescriptions assigned to the infant subgroup were for patients who were aged 1 year at the time of fills, introducing some degree of misclassification bias.

In this national interrupted time series analysis, acid blocker dispensing to US children declined sharply after the ranitidine recall, illustrating the power of drug recalls to affect prescribing decisions in children. After the COVID-19 pandemic began, acid blocker dispensing recovered to a greater degree in older children compared to infants. Future research should investigate the reasons for these diverging trends.

H2RA

H2 receptor antagonists

PPI

proton pump inhibitors