Steroids for Pediatric Adenotonsillectomy Pain—A Prospective Multicenter Randomized Trial

Abstract

Objective

To compare postoperative caregiver‐reported pain control of two steroid regimens following pediatric adenotonsillectomy. Secondary objectives assessed differences in analgesic use, return to normal diet, caregiver calls, adverse effects, and emergency room (ER) visits.

Study Design

Prospective randomized pragmatic trial.

Setting

Three academic tertiary care children's hospitals.

Methods

Healthy children aged 3 to 10 undergoing adenotonsillectomy were recruited and randomly assigned to receive either 0.6 mg/kg dexamethasone on postoperative day 3 or prednisolone 0.5 mg/kg days 1 to 3. Both groups continued standard weight‐based regimens of acetaminophen and ibuprofen. A blinded research team member performed postoperative telephone surveys using the Wong‐Baker FACES pain‐rating scale. Secondary outcomes were reviewed using a combination of a telephone survey and electronic health record review.

Results

A total of 119 patients were randomly assigned (61 [51%] female, 58 [49%] male, and mean age of 5.8 years (95% CI, 5.5‐6.2), with 59 (49.6%) in the dexamethasone group and 60 (50.4%) in the prednisolone group. The pain score was lower in the prednisolone group (4.0 [3.7‐4.3]) in the univariate model. However adjustment for covariates and random effects mitigated this difference (b = 0.00; 95% CI, −0.52 to 0.52; P = .99). There were no significant differences in analgesic use, return to normal diet, calls, adverse effects, or ER visits. There were no subgroup effects related to surgical technique.

Conclusion

Prednisolone and dexamethasone‐based regimens show comparable efficacy for pain scores, return to diet and resumption of activities with no reported adverse events. Shared decision‐making should involve caregivers to select the regimen best suited for their child's circumstances.

Adenotonsillectomy (AT) remains the most common pediatric ambulatory surgical procedure in the United States. In 2019, more than 460,000 children younger than 17 underwent AT in hospital‐owned facilities. ¹ This procedure is generally well‐tolerated; most children are discharged home on the same day. Unanticipated postoperative emergency department (ED) visits range from 6 % to 11% and are due to postoperative pain, nausea, emesis, and dehydration. ² , ³ Poorly controlled pain may decrease oral intake, leading to dehydration, and pain management is understandably a significant parental concern. ⁴ , ⁵ The American Academy of Otolaryngology–Head and Neck Surgery clinical practice guidelines support using intraoperative intravenous dexamethasone and postoperative ibuprofen and acetaminophen for improved recovery. These guidelines also include perioperative caregiver pain counseling, further emphasizing the importance of adequate pain management. ⁶

Post‐AT pain lasts a median of 9 days, peaks between 3 and 5 days following surgery, and is associated with increased difficulty eating and sleeping. ⁷ Most unplanned ED visits for uncontrolled pain occur on postoperative day 4. ⁸ With dexamethasone already recommended for intraoperative use, this postoperative pain peak supports the potential utility of additional corticosteroids in the management of post‐AT pain. Postoperative steroids improve pain, diet, activity, and reepithelialization. ⁹ They also decrease nausea and vomiting ¹⁰ and phone calls for pain‐related concerns. ¹¹ Notably, a previous study from our institution showed that a single dose of dexamethasone administered on postoperative day 3 improved pain and decreased analgesic use, as well as phone calls, compared to children who received only analgesics. ¹²

The rationale for comparing prednisolone and dexamethasone lies in their similar pharmacological properties and potential clinical benefits, while having different durations of action and hence dosing requirements. Dexamethasone has a longer biological half‐life (36‐72 hours), which could provide extended anti‐inflammatory effects with a single dose than prednisolone (12‐36 hours) that needs additional doses for physiological comparability. ¹³ , ¹⁴ However, prednisolone is more widely available in liquid form, improving compliance in children. ¹⁵ Both steroids, despite having similar anti‐inflammatory potency, could have different effects in managing post‐AT pain due to their pharmacokinetics. ¹⁶

The primary objective of this study was to compare two different oral steroid regimens in children undergoing AT to refine the type, dose, and duration of optimal therapy. We hypothesized that by comparing outcomes of administering dexamethasone on day 3 and prednisolone on days 1 to 3, the steroid with superior efficacy and fewer side effects could be identified with improved pain control, faster return to regular diet and activity, as well as fewer caregiver‐initiated phone calls and ED visits.

Methods

Patients

Children with sleep‐disordered breathing aged 3 to 10 years who were scheduled to undergo AT from October 2021 to September 2023 were prospectively recruited as part of this pragmatic study at three tertiary care academic children's hospitals for the study. Children with significant comorbidities, including cardiopulmonary disease, sickle cell anemia, poorly controlled asthma, Down syndrome, and craniofacial disorders, neurocognitive issues, recurrent or chronic tonsillitis, or chronic steroid therapy were excluded. Only English‐speaking patients and families were included in the study due to the reliance on telephone follow‐up surveys and standardized instructions for postoperative care and medication administration. The study protocol, surveys, and consent forms were approved by the institutional review board of the University of Maryland, Baltimore, and each participating institution (Children's National Hospital, Washington, DC, Texas Children's Hospital, Houston, TX). Consent was obtained from all children's primary caregivers. Assent was not obtained from the children given (i) the relatively younger inclusion age, and (ii) outcome data were collected from caregivers.

Patients were randomly assigned into two groups: (1) dexamethasone group to receive one oral dexamethasone dose 0.6 mg/kg (up to 8 mg) on postoperative day 3 and (2) prednisolone group to receive oral prednisolone dose of 0.5 mg/kg (up to 30 mg) on postoperative days 1 to 3. As medication dispensing varied across the three participating institutions, the study protocol allowed for either direct dispensing of study medications (dexamethasone or prednisolone) by the hospital pharmacy before discharge or prescriptions for caregivers to obtain from community pharmacies. In both scenarios, detailed instructions were provided to caregivers. Medication acquisition was verified during the scheduled telephone assessments on days 4 and 7.

Oral dexamethasone was administered either in liquid or tablet form, with the latter needing crushing and dissolution in a liquid before administration. There was no financial compensation and no reimbursement for any medication. Enrollment and allocation are depicted in the Consolidate Standards of Reporting Trials (CONSORT) diagram in Figure 1. ¹⁷ Additionally, both groups received a standard analgesic regimen of alternating weight‐based doses of ibuprofen and acetaminophen. Specifically, acetaminophen was prescribed at 10 to 15 mg/kg every 4 to 6 hours (not exceeding 5 doses in 24 hours) and ibuprofen at 5 to 10 mg/kg every 6 to 8 hours (not exceeding 4 doses in 24 hours).

Figure 1.

Consolidate Standards of Reporting Trials flow diagram. A diagram illustrating the flow of participants through the trial. No patients were lost to follow‐up or discontinued the intervention. All patients were included in the final analysis.

Randomization was performed by predetermined block randomization. Both intracapsular and extracapsular tonsillectomy techniques were permitted in this study, given the pragmatic trial design. The selection of surgical technique was based on the surgeon's clinical judgment and standard practice at each institution rather than being mandated by the study protocol. Procedures were performed by a combination of attending surgeons, fellows, and residents under supervision, reflecting the real‐world clinical environment across the participating academic institutions. This pragmatic approach aimed to capture the spectrum of surgical expertise typically involved. A standardized anesthetic technique was used for all patients. ¹⁸ At induction of anesthesia, patients received a single intravenous dose of dexamethasone (0.15 mg/kg) as part of the routine practice at all institutions with no additional intravenous medications beyond a standard anesthetic regimen. This study focused exclusively on ambulatory AT cases reflecting typical patient care patterns encountered by otolaryngologists in the community. Caregivers were given a general postoperative instruction sheet at discharge, including diet and activity recommendations. Standard postoperative instructions were provided in print form to all patients at discharge. These instructions included specific guidance on diet progression (starting with clear liquids, advancing to soft foods, and gradually returning to normal diet as tolerated), activity restrictions (avoiding strenuous physical activity, contact sports, and heavy lifting for the first week), and hydration requirements. This standardization of postoperative care across both treatment groups controlled for the potential confounding effect of varying recovery practices on pain outcomes. Adherence to these instructions was reinforced during follow‐up telephone calls.

Sample Size Calculation

A priori sample size calculation was performed for a two‐group comparison using a two‐tailed, two‐sample t test. Assuming a medium effect size (Cohen's d = 0.5), α = .05, β = .80, the estimated required sample size was 60 participants/group (total = 120) to detect a statistically significant difference between the groups. This sample size should provide adequate power to detect a medium‐sized effect while balancing the risk of type I/II errors.

Outcome Measures

On the day of surgery, patients and families were provided with a postoperative survey (Supplemental Figure S1, available online). On postoperative days 4 and 7, a research nurse blinded to the study condition called each patient to review the survey. These specific timepoints were selected based on our previous research demonstrating that post‐AT pain typically peaks between days 3 and 5 and begins significant resolution by day 7. ¹² These timepoints capture the critical window during which intervention effects would be most pronounced and clinically relevant. The primary pain outcome was assessed using the well‐validated visual analog Wong‐Baker FACES pain‐rating scale. ¹⁹ , ²⁰ Secondary outcomes included the number of postoperative phone calls, returns to the ED, and the number of days to return to normal diet and activity as described by the caregiver and further review of the electronic health record. Research nurses specifically inquired about any unexpected symptoms or complications in addition to our primary and secondary outcome measures. Caregivers were explicitly instructed to report any unusual symptoms or concerns.

Statistical Analysis

Categorical variables were described by n (%) and continuous variables by mean (95% confidence interval). Baseline characteristics were compared between groups using a χ ² test. The distribution of the outcomes between groups was visualized by boxplots. The primary and secondary outcomes (continuous variables) were compared between groups using linear mixed‐effects regression models implemented using R (version 4.3, https://cran.r-project.org). ²¹ For example, the linear mixed‐effects model for pain score as the main outcome was compared between treatment arms and adjusted for age, sex assigned at birth, race, type of surgery (intracapsular vs extracapsular), and the postoperative day for the reported outcome. Random effects included participant and patient ID. We expected that surgery type as a covariate in our linear mixed‐effects model statistically controlled for the expected effect of surgical technique on pain outcomes. Additionally, we conducted a subgroup analysis for patients who underwent extracapsular tonsillectomy to validate the findings across procedural strata. Mixed‐effects models incorporate all data from study participants at all sites with at least one measurement across timepoints and account for missing data using a maximum‐likelihood‐based approach. ²² , ²³ All tests were two‐sided at α = .05.

Results

A total of 119 children were recruited from the three sites. All patients in both treatment groups were discharged on the day of surgery. No overnight admissions or additional medications were required. All study participants successfully obtained their assigned medications through either hospital or community pharmacies. There were no reported delays in medication acquisition that affected the timing of the first dose administration for either treatment group. No adverse events specifically attributable to either steroid regimen were reported throughout the study period. The mean age was 5.8 years (95% CI, 5.5‐6.2), and 61 were female (51%). The racial distribution was black = 42 (35%), white = 40 (34%), and other = 37 (31%). The distribution of demographic characteristics was balanced between the groups (Table 1).

Table 1.

Comparison of Demographic Characteristics, Primary, and Secondary Outcomes ^a

	Dexamethasone	Prednisolone
Age, mo	68.9 (62.5‐75.4)	70.9 (65.4‐76.4)
Sex
Male	29 (49.2%)	29 (48.3%)
Female	30 (50.8%)	31 (51.7%)
Race
Black	18 (31.6%)	21 (35.6%)
White	22 (38.6%)	20 (33.9%)
Asian	4 (7%)	3 (5.1%)
Other	13 (22.8%)	15 (25.6%)
Surgery type
Extracapsular	50 (84.8%)	50 (83.3%)
Intracapsular	9 (15.3%)	10 (16.7%)
Pain score	4.3 (4.0‐4.6)	4.0 (3.7‐4.3)
ED visits	3 (5.1%)	2 (3.3%)
Phone calls	12 (20.3%)	9 (15%)
Return to normal diet	5.5 (4.9‐6.1)	5.5 (4.9‐6.1)
Return to normal activity	5.4 (4.8‐6.0)	5.8 (5.2‐6.5)

Abbreviation: ED, emergency department.

^a Categorical variables are shown as n (%), and continuous variables are shown by mean (95% confidence interval).

Figure 2A shows the unadjusted pain scores grouped by the treatment arm and stratified by the postoperative day. Univariate models showed superior outcomes for pain score (average pain score, 4.0; 95% CI, 3.7‐4.3) for the prednisolone group compared to the dexamethasone group (average pain score, 4.3; 95% CI, 4.0‐4.6). However, adjustment for the covariates and random effects mitigated the differences. Fitted linear mixed‐effects models (Table 2) demonstrated that the postrandomization pain scores were not different between study arms (b = −0.37; 95% CI, −1.32 to 0.58; P = .45). Additional associations were identified for the type of surgery (β = −1.36; 95% CI, −2.10 to −0.63, P < .001) and postoperative day (β = −0.32; 95% CI, −0.41 to −0.23; P < .001). These coefficients demonstrate that pain scores were lower in children who underwent the intracapsular technique and decreased over time in all children. Given the impact of surgical technique, a subgroup analysis related to children who underwent extracapsular tonsillectomy also did not demonstrate the difference in pain scores between the trial arms (Table 3, b = 0.24; 95% CI, −0.33 to 0.81; P = .41). The adjusted pain scores are shown in Figure 2B. Other secondary outcomes—return to regular diet (β = −0.10; 95% CI, −0.89 to 0.68; P = .80) and activity (β = −0.46; 95% CI, −0.45 to 1.38; P = .32) were also unaffected by the type of steroid administered. Lastly, there were no statistically significant differences in the number of ED visits (3 in dexamethasone vs 2 in prednisolone, χ ² = 0.0, P = .68) or phone calls made to the office (12 in dexamethasone vs 9 in the prednisolone group, χ ² = 1.4, P = .53). Pain management and reduced oral intake were the predominant concerns in both groups, with no significant between‐group differences. No other adverse events were reported in the trial arms.

Figure 2.

Pain scores. (A) Unadjusted scores boxplot and (B) pain scores (confidence intervals), stratified by group and day, and adjusted for age, sex, race, surgery type, day, and random effects of subject IDs.

Table 2.

Multivariable Regression Analysis of Pain Scores Showing Fixed Effects (Sex, Race, Day, and Surgery Type), Random Effects (Patient/Site), and Model Fit Statistics With Coefficients and Significance Levels

Predictors	Estimates	CI	P
(Intercept)	6.74***	5.30‐8.18	<.001
Age	0.00	−0.02 to 0.01	.399
Sex	−0.30	−0.82 to 0.22	.261
Race: black	−0.95	−1.98 to 0.09	.073
Race: white	−0.67	−1.61 to 0.28	.166
Race: other	−1.33	−3.47 to 0.81	.224
Surgery type	−1.36***	−2.10 to −0.63	<.001
Group	0.00	−0.52 to 0.52	.996
Day	−0.32***	−0.41 to −0.23	<.001
Random effects
Variance (within subjects)	7.36
Variance (between subjects)	3.40
Variance (between sites)	0.32
Intraclass correlation coefficient	0.60
Observations	833
Marginal R 2/conditional R 2	0.079/0.370

^*** P < 0.001.

Table 3.

Multivariable Regression Analysis of Differences in Pain Scores in the Subset of Children Who Underwent Conventional Extracapsular Tonsillectomy Showing Fixed Effects (Sex, Race, Day, and Surgery Type), Random Effects (Patient/Site), and Model Fit Statistics With Coefficients and Significance Levels

Predictors	Estimates	CI	P
(Intercept)	7.29***	5.61‐8.97	<.001
Age	0.00	−0.01 to 0.01	.823
Sex	−0.62*	−1.21 to −0.02	.041
Race: black	−2.12**	−3.41 to −0.84	.001
Race: white	−1.51*	−2.82 to −0.19	.025
Race: other	−1.49	−2.97 to −0.01	.049
Group	0.24	−0.33 to 0.81	.406
Day	−0.30***	−0.41 to −0.20	<.001
Random effects
Variance (within subjects)	7.48
Variance (between subjects)	2.73
Variance (between sites)	0.22
Intraclass correlation coefficient	0.28
Observations	700
Marginal R 2/conditional R 2	0.034/0.153

^* P < 0.05;

^** P < 0.01.

^*** P < 0.001.

Discussion

AT is among the most common and one of the most painful pediatric surgical procedures. Postoperative pain is reported on average for 9 &days and may last up to 2 weeks. ⁷ Different AT techniques exist, and the intracapsular method is associated with less postoperative pain than the extracapsular method. ²⁴ , ²⁵ , ²⁶ , ²⁷ Poorly controlled pain leads to reduced oral intake with subsequent dehydration, restless sleep, behavioral changes, nausea, and vomiting and may result in unplanned ED visits. ² , ³ , ⁸ , ²⁸ Billings et al demonstrated that 35% of all children returning to the ED with postoperative pain had undergone AT. ⁸ In their analysis of 139,715 pediatric AT patients, Mahant et al found that 2.2% presented for vomiting and dehydration. ² Similarly, when considering ED visits following AT, 7.6% of all revisits were related to dehydration, whereas 2.3% were related to nausea and vomiting. ³

In procedure‐specific pediatric studies, postoperative pain following AT was shown to be bimodal. It initially peaks on the day after the procedure, with a second peak between days 3 and 5. ⁷ , ²⁹ Although various analgesic regimens have been studied, doses of acetaminophen (10 mg/kg) and ibuprofen (5 mg/kg) every 3 hours effectively controlled post‐AT pain in most children without an increase in bleeding. ³⁰ This contributed to the most recent strong recommendation by the AAO‐HNS to use ibuprofen, acetaminophen, or both for pain control following AT. ⁶

Although commonly used in adults postoperatively, opioids are avoided in children due to the associated side effects of respiratory depression, nausea/vomiting, urinary retention, and sedation. ⁶ , ³¹ Children undergoing AT for obstructive sleep apnea are at higher risk as the associated respiratory depression may be fatal. Notably, in 2013, reports of respiratory depression and pediatric deaths following AT in patients treated with codeine led the Food and Drug Administration to issue a boxed warning against the use of products containing the opioid codeine. ³² Codeine is converted to morphine by the liver, and children who are genetically predisposed to ultra‐rapid metabolism of the drug can accumulate life‐threatening or fatal amounts of morphine in the body. Children with obstructive sleep apnea who received codeine for pain relief following a tonsillectomy are particularly vulnerable to this effect. Nonsteroidal anti‐inflammatory drugs (NSAIDs) are a safe nonopioid alternative for postoperative pain management. When compared with placebo or opioids following AT, NSAIDs use provided adequate analgesia without increasing the risk of bleeding. ³³ , ³⁴ Given the risks of opioid‐induced apnea, we sought to optimize a multimodal analgesia regimen with the use of postoperative steroids.

Systemic corticosteroids are widely used for postoperative pain management, and their use specifically in AT has been studied for more than 50 years. ²⁸ , ³⁵ Any surgical procedure, including AT, leads to mechanical tissue damage and distortion of nerve terminals, with a subsequent tissue inflammatory response that includes bradykinin release and nociceptive pain receptor activation. ²⁸ , ³⁶ Corticosteroid effects are primarily anti‐inflammatory. They suppress bradykinin release, reduce inflammatory mediators, and inhibit cyclooxygenase and lipoxygenase pathways, which reduces prostaglandin production. ²⁸ , ³⁷ , ³⁸ , ³⁹ Currently recommended regimens for postoperative AT pain control include the use of ibuprofen and acetaminophen targeting only the cyclooxygenase and lipoxygenase pathways, thus addressing only a portion of the inflammatory response. ⁶

Compared to NSAIDS, glucocorticoids exhibit superior anti‐inflammatory efficacy. A wide array of postoperative corticosteroid regimens have been studied in both adults and children undergoing AT, with benefits seen in children given three postoperative doses of dexamethasone (0.5 mg/kg) ¹¹ or a 7‐day course of prednisolone (0.25 mg/kg/day). ⁹ The postoperative administration of three doses of 0.5 mg/kg dexamethasone was shown to decrease postoperative parental phone calls by 9% and the risk of postoperative AT hemorrhage by 7% in a retrospective review of 1200 children. ¹¹ At a dose of 0.25 mg/kg/day, a 7‐day course of prednisolone showed no significant benefits on postoperative day 1. However, the regimen was associated with decreased pain scores on day 7, along with improvement in diet, activity, reepithelialization on endoscopic examination, fever, and sleep disturbance. ⁹

Accounting for the known second peak in postoperative pain, we previously demonstrated improved pain control by administering a single dose of dexamethasone (0.6 mg/kg) on postoperative day 3. ¹² With this multi‐institutional study, we compared the dexamethasone regimen to a prednisolone regimen (0.5 mg/kg) regimen on postoperative days 1 to 3. The generally used therapeutic dose of prednisolone is 1 mg/kg/day, with a range from 0.14 to 2 mg/kg/day, depending on the situation. The various glucocorticoids are equivalent in anti‐inflammatory efficacy and have similar side effect profiles, except for fluid retention. Everyone received the standard dose of intraoperative dexamethasone. The dexamethasone dose of 0.6 mg/kg was chosen due to its well‐known anti‐inflammatory action with improved outcomes in patients with croup. ⁴⁰ We decided on 0.5 mg/kg/day of prednisolone as the dose that would be best suited to our current study population in terms of lower risk of side effects while maintaining a duration of anti‐inflammatory effect similar to dexamethasone. Our study did not reveal any differences between the prednisolone and dexamethasone‐based regimens in the studied outcomes of average pain scores, return to normal diet, and resumption of regular activities with no reported medication‐related adverse events.

Although ethically justified, the absence of a placebo arm limits our ability to assess the absolute effect of steroids. Although nursing staff were blinded, caregivers were not, potentially introducing bias, though we believe this impact was minimal. Oral dexamethasone was administered either in liquid form, which is not always available depending on the location, or as a tablet that needed to be crushed and dissolved for administration. This can affect compliance and accuracy of dosing. Additionally, the cost difference between the two drugs that varies regionally, based on insurance and pharmacy type, as well as the duration of therapy, could also impact patient preference and dosing reliability. Although we provided standardized postoperative instructions to all patients, individual variation in adherence to these guidelines regarding diet progression and activity restrictions may have influenced pain outcomes. However, our randomized design ensured that any such variations would be distributed equally between treatment groups, allowing for valid comparison of the steroid regimens. Additionally, our secondary outcomes measuring return to normal diet and activity provide insight into functional recovery patterns that align with our pain score findings. Our findings may also be limited in generalizability across various AT methods.

Our study focused on specific steroid regimens, potentially overlooking other effective dosing strategies or steroid types. Our focus on the immediate postoperative period also precluded assessment of long‐term outcomes or delayed side effects. Although we acknowledge that some patients may experience prolonged pain beyond our assessment period, our primary aim was to evaluate the comparative efficacy of steroid regimens during the acute postoperative period rather than extended recovery. The inclusion of multiple institutions and surgeons with varying levels of experience strengthens the external validity of our findings while addressing the potential variability through mixed‐effects modeling with site as a random effect. The comparable outcomes between treatment groups, despite this procedural heterogeneity, suggest that the efficacy of these steroid regimens remains consistent across different clinical settings and surgeon experience levels—an important consideration for broad clinical implementation.

Despite limitations, the multi‐institutional design and use of a previously validated control arm strengthen our findings, providing valuable insights into postoperative pain management for pediatric AT patients. By demonstrating comparable efficacy between a single‐dose dexamethasone regimen and a 3‐day prednisolone course, our results provide clinicians with evidence‐based flexibility in their postoperative steroid prescribing. For example, prednisolone is more widely available in liquid form, potentially improving administration compliance in younger children, whereas the single‐dose dexamethasone approach may be preferable for families concerned about medication adherence over multiple days. The choice between these regimens can be personalized based on patient and family preferences, medication access, and institutional formulary considerations.

Future studies should address these limitations by incorporating a wider range of surgical techniques, exploring additional steroid regimens, newer and safer anti‐inflammatory and analgesic drugs, controlling for additional variables, and with longer‐term follow‐up until complete resolution of pain with return to normal diet and activity.

Conclusions

Our findings reveal comparable efficacy between prednisolone and dexamethasone‐based regimens regarding average pain scores, return to normal diet, and resumption of regular activities. Both steroid regimens exhibited favorable safety profiles, with no adverse events reported throughout the study. The choice between prednisolone and dexamethasone may be influenced by factors such as medication availability, ease of administration, and individual patient characteristics. We recommend involving caregivers in shared decision‐making to select the regimen best suited to their child's circumstances. This approach allows personalized care while maintaining effective pain control and minimizing potential risks. Future research should focus on refining steroid dosing, newer analgesics, and anti‐inflammatory protocols with good safety profiles and early return to normal function.

Author Contributions

Kevin D. Pereira, study design, data acquisition and analysis, manuscript writing, and revision of manuscript, presentation of research, and final approval; Wiktoria A. Gocal, data analysis, conduct, manuscript draft and revisions, presentation of research; Anna V. Borodianski, study conduct, data acquisition and analysis, presentation, manuscript drafting; Mary E. Williamson, study design and conduct, data acquisition, presentation, manuscript drafting; Hengameh K. Behzadpour, conduct of research, data acquisition, analysis, manuscript review and editing; Sonal Saraiya, study design, conduct, data acquisition and analysis, manuscript drafting, editing, and final approval; Diego A. Preciado, study design, conduct, data analysis, manuscript drafting, editing, and final approval; Amal Isaiah, study design, data analysis and interpretation, manuscript draft, revision and final approval.

Disclosures

Competing interests

None.

Funding source

None.

Supporting information

Supporting information.