Sage Journals: Discover world-class research

Abstract

Purpose:

In 2005, the Food and Drug Administration (FDA) issued a decision memorandum regarding nonsteroidal anti-inflammatory drugs (NSAIDs). The memorandum recommended the withdrawal of certain NSAIDs due to potential cardiovascular adverse effects. It highlighted the issue of cardiovascular risk associated with NSAIDs as a class. The NSAID medication guide includes a wide range of adverse drug reactions (ADRs), such as increased blood pressure, liver failure, allergic reactions, heart attack, and intestinal bleeding. Although both sexes have an increased risk of ADRs with NSAID use, females have a greater risk than males due to differences in pharmacodynamics and higher medication concentrations (mg/kg). In particular, females with high blood pressure, coronary heart, kidney, and liver disease are at an additional risk of harm from NSAID ADRs. This study quantifies sex-specific differences and other factors associated with prescription NSAID use.

Method:

The data for this study were obtained from the National Health and Nutrition Examination Survey (NHANES), a complex survey conducted by the Centers for Disease Control and Prevention (CDC) in two-year cycles. A survey-weighted logistic regression model was utilized to investigate potential sex differences with prescription NSAIDs in the context of other factors, including kidney disease, hypertension, liver disease, insurance status, coronary heart disease, and age, within the 2011–2018 NHANES survey data.

Results:

Females reported a slightly higher percentage of high blood pressure and kidney disease than males, while males reported a slightly higher percentage of coronary heart and liver disease than females. Last, the model indicated that females were 58% more likely to have used a prescription NSAID than males.

Conclusion:

The results confirm that women and people with medical conditions, who would potentially suffer greater harm from NSAID ADRs, are more likely to use a prescription NSAID than individuals without these conditions.

Introduction

Nonsteroidal anti-inflammatory drugs (NSAIDs) are widely used in the United States, with 90% of NSAID prescriptions held by individuals aged 65 years or older.^1,2 Furthermore, the prevalence of NSAID use worldwide among people aged 65 years and older, with at least one present NSAID prescription, ranges from 10% to 20%. Prescription NSAIDs are generally more potent targeting inflammation and pain compared with their over-the-counter (OTC) counterparts, but they also come with greater risks. These specific risk factors include certain medical conditions, female sex, age, and racial group. Due to the prevalence of prescription NSAID use, this topic is deemed a significant public health issue.

In 2005, the Food and Drug Administration (FDA) issued a decision memorandum regarding NSAIDs. The memorandum recommended the withdrawal of certain NSAIDs due to potential cardiovascular adverse effects. It highlighted the issue of cardiovascular risk associated with NSAIDs as a class, with the exception of aspirin. Aspirin was noted to have beneficial pharmacological properties that set it apart from other NSAID medications.^3,4 The decision memorandum by the FDA outlined the possible side effects of NSAIDs. Furthermore, the FDA advised that all prescription NSAIDs be accompanied by a medication guide warning patients of the increased risk of heart attacks, strokes, and stomach bleeding.^3,4 There is also evidence that suggests NSAIDs can harm the kidneys and liver. Consequently, prescription NSAIDs are generally contraindicated for individuals with hypertension, heart failure, and kidney or liver disease. This study investigates the use of prescription NSAIDs in individuals with these conditions, placing them at increased risk of harm from prescription NSAID adverse drug reactions (ADRs). In addition, this study examines how demographic factors, such as age, ethnicity, insurance status, and sex, are associated with prescription NSAID use. Specifically, the study focuses on the differences in prescription NSAID use between males and females, as females have been reported to be more susceptible to NSAID ADRs.^5,6 These sex-specific differences will be interpreted in the context of these other relevant factors.

The medication guide for prescription NSAIDs lists several ADRs, such as increased blood pressure, liver failure, allergic reactions, heart attack, and intestinal bleeding.⁷ Although both sexes are at an increased risk of experiencing ADRs with NSAID use, females have a higher risk than males due to differences in pharmacodynamics and higher medication concentrations (mg/kg). For instance, men are at a 1.4× greater risk of experiencing serious gastrointestinal (GI) issues than women, regardless of NSAID use. However, women are more susceptible to NSAID-related GI events.⁸ Studies have shown sex differences in liver injury, GI problems, hypertension, and skin rashes with NSAIDs.^5,8

–12 The Nurses’ Health Study, which included more than 80,000 women, discovered a strong link between the development of hypertension and NSAID use.^10,13 In addition, NSAID use has been associated with miscarriage and premature birth.^14,15

Research studies have shown that women are more likely than men to receive prescriptions for NSAIDs.^5,16,17 However, the FDA’s decision memorandum regarding NSAIDs does not mention any differences by sex.^3,4 This may be due to the limited availability of sex-specific data on the effects of prescription NSAIDs. Clinical trials evaluating prescription NSAIDs often overlook sex-specific differences and underrepresent women.^16,18,19 This study addresses a present public health concern by examining sex-specific differences in ADR risk regarding NSAID usage, an area frequently neglected in clinical trials.^16,18 Although the FDA’s 2005 decision memorandum on NSAIDs primarily focused on cardiovascular issues, this research seeks to fill a knowledge gap by investigating sex differences in ADRs related to specific prescription NSAIDs, an area that has received limited research attention.³

This study was intended to examine the present use of NSAIDs in female subjects who are at additional risk of harm from NSAID ADRs producing meaningful risk estimates by modeling data from the National Health and Nutrition Examination Survey (NHANES) to make inferences applicable to the Nation. This is not the first study to analyze and model NSAID use with NHANES. Prior research that modeled NSAID use with NHANES focused primarily on demographics and the overall prevalence restricting their analysis to a comparison of the 1988 through 1994 with the 1999 though 2004 NHANES cycles.²⁰

Methods

Data source

The data for this study were obtained from NHANES, a complex survey conducted by the Centers for Disease Control and Prevention (CDC) in two-year cycles. NHANES is designed to make inferences about the health of the U.S. civilian noninstitutionalized population. Part of the survey involves a trained individual administering an in-home questionnaire collecting general information on health, demographics, and medication usage. The participants reported their prescription medication usage over the past 30 days, with the interviewers verifying the medication and dose by examining the participant’s prescription bottles. Our analysis comprised data from four NHANES cycles, which included participants from the years 2011 to 2018, aged 20 years and older. After combining the 2011–2018 NHANES data files, we adjusted the sampling weights according to the National Center for Health Statistics guidelines.²¹

The questionnaire used in the study addressed several broadly defined medical conditions, such as high blood pressure, coronary heart disease, kidney disease, and liver disease, which place an individual at greater risk of the harmful effects of NSAID ADRs. To determine the presence or absence of these conditions, we relied on NHANES responses to questions BPQ020 (hypertension), KIQ022 (weak or failing kidneys), MCQ160I (liver condition), and multiple questions (MCQ160C, MCQ160B, MCQ160E, and MCQ160D) for coronary heart disease (CHD). If any of the four CHD questions had a positive response, CHD was recorded as present; otherwise, it was recorded as negative or unknown. In addition, we examined the association between insurance coverage (HIQ011) and prescription NSAID use.

Survey-weighted logistic regression

A survey-weighted logistic regression model was used to investigate potential sex differences in the relationship between prescription NSAIDs, kidney disease, hypertension, liver disease, insurance status, coronary heart disease, and age. Equation 1 represents this model, which generates odds ratios as the risk measure. The independent variables were coded as categorical variables, facilitating model computations and comparisons within combinations of categories relative to the dependent variable. This modeling approach, with prescription NSAID use as the dependent variable, enables us to discern differences within the categories of the independent terms, including sex. In addition, the model allows the calculation of probabilities for prescription NSAID use leveraging combinations of the independent variable categories. These probabilities allow us to estimate the number of people at the national level who fall within these groupings of independent variable categories. Hypertension, liver disease, insurance status, and CHD were coded as binary variables, with one indicating the presence of the condition and a zero as the reference value. Sex was also coded as a binary variable, with males and females coded as zero and one, respectively. Race and ethnicity were coded such that the non-Hispanic white category was a reference for the remaining racial and ethnic groups. Although age was provided as a continuous variable, we created 10-year age categories ranging from 20 to 80 plus years of age. NHANES reports the age of subjects 80 years and older simply as 80 plus years, which prevents knowing the exact ages of more elderly participants. The 20- to 30-years-age category was coded as the older groups’ reference group.

\begin{array}{l} \ln {\frac{\Pr (Y = NSAID)}{\Pr (Y = N o NSAID)}} = β_{0} + β_{1} [Sex] + β_{2} [Hypertension] + β_{3} [Kidney] + β_{4} [Race / Ethnicity] + β_{5} [Insurance] + β_{6} [CHD] + β_{7} [Age] + β_{8} [Liver] \end{array}

(1)

The list of prescription NSAIDs examined in this study includes some commonly used medications such as ibuprofen, naproxen, and diclofenac, as well as less well-known drugs such as diflunisal and meclofenamate sodium. A complete list of prescription NSAIDs examined in this study is found in Supplementary Appendix Table S1. The NHANES questionnaire surveys the participant’s use of prescription NSAIDs over the previous 30 days. It is also important to note that only prescription NSAIDs were examined, and OTC medications, which are widely used and may contribute to overall NSAID exposure, were not considered. Many prescription NSAIDs have an OTC counterpart leading to concerns of misclassification for NSAID exposure. Prior research has indicated that prescription NSAID claims data can give valid estimates of association even when an OTC version is available when OTC use is relatively high compared with prescription drug use.²² Yood et al. (2007) concluded “that under realistic conditions of NSAID exposure or other drugs with similar prescription and OTC use patterns, using claims data without information on OTC drug use may have a small, or even negligible effect on validity” (p. 966). In our sample of adults aged 20 years and older, 5.9% were surveyed as having used a prescription NSAID within the last 30 days. Sales of NSAID OTC medications in the United States are difficult to track, but estimates are high as these drugs account for 60% of the total analgesic market in 2016.²³

Weight adjustment for confounding

NHANES is frequently used to support studies on the national prevalence of health conditions and prescription drug use. Although a complex survey, it is still an observational study subject to the same limitations, such as confounding, that plague cross-sectional designs. While NHANES data are commonly used for calculating prevalences, more complex analyses such as regression that address confounding are rare. One way to address confounding is through propensity scores, which are widely used in conventional observational studies. Propensity scores help balance the systematic differences between male and female survey participants who use prescription NSAIDs and those who do not, Equation 1. The use of propensity scores can help achieve balance between groups based on observed factors, but it may not extend to unmeasured or unknown covariates.²⁴ The propensity scoring approach in our survey-weighted logistic regression model utilized sex as the treatment variable while providing for a balance of systemic differences within the remaining original independent variables from Equation 1. Unfortunately, propensity score methods are not fully integrated into software designed for complex survey analysis.

Propensity scores can be challenging to use with complex survey data due to the need to account for complex survey design variables, such as sampling weights, to make inferences about the U.S. population. To address this issue, we adopted the approach by DuGoff et al. who essentially recommended multiplying the sampling weights with the propensity score weights. By doing so, we can effectively control for confounding and maintain the ability to draw inferences about the U.S. population.²⁵ This study is not the first to utilize DuGoff’s method to adjust for confounding in NHANES data.²⁶ The propensity score model used in this study is designed to produce average treatment effect (ATE) weights.

The ATE refers to the overall effect of prescription NSAIDs on all study participants, regardless of whether they were exposed to the medication. The ATE represents the difference in outcomes between all subjects in the study if they had taken prescription NSAIDs versus if they had not taken them.²⁷ When all study subjects are potentially exposed to prescription NSAIDs, the ATE may be the most appropriate weighting choice.²⁸ In this study, we generated ATE weights through our propensity score model and multiplied them in combination with our survey sampling weights to adjust for confounding in our survey-weighted logistic regression model.

All statistical analyses were conducted using the survey procedures available in SAS version 9.4 and was considered statistically significant at an alpha of five percent.

Results

Descriptive measures

Supplementary Appendix Tables S2–S4 provide descriptive statistics and frequencies for the factors included in the model. The average age of individuals who used NSAIDs within the past 30 days was higher than those who did not, for both males and females. The surveyed population was predominantly non-Hispanic white. Females reported a slightly higher percentage of high blood pressure and kidney disease than males, while males reported a slightly higher percentage of coronary heart and liver disease than females. Furthermore, a larger percentage of females reported having health insurance coverage than males.

Survey-weighted logistic regression risk estimates

Table 1 presents our survey-weighted logistic regression analysis results in terms of the odds ratio. Our findings indicate that females are nearly 60% more likely than males to have used a prescription NSAID within the previous month. Individuals with a history of high blood pressure are almost 82% more likely to have used a prescription NSAID than those without this condition. Regarding race/ethnicity, those who reported other or mixed race were about 56% more likely to have used a prescription NSAID than non-Hispanic whites, while non-Hispanic Asians were 49% less likely. Prescription NSAID users were 62% more likely to have reported having insurance than nonusers. Individuals with a history of liver disease were approximately 48% more likely than those without this condition to have used a prescription NSAID within the last month. The results for coronary heart and kidney disease were not statistically significant.

Table 1.

Survey-Weighted Logistic Regression Results for Model Factors in Terms of Odd Ratios

Odds ratio estimates
Parameter	Point Estimate	95% Confidence Limits
Sex (Male is Reference)	1.583	1.345 1.864
Hypertension (No is Reference)
Yes	1.816	1.507 2.187
^*Do not know	1.912	0.389 9.399
Race/Ethnicity (White is Reference)
^*Mexican American	1.161	0.917 1.471
Non-Hispanic Asian	0.511	0.389 0.671
^*Non-Hispanic black	1.024	0.857 1.223
^*Other Hispanic	1.090	0.869 1.368
Other race—including multiracial	1.562	1.142 2.137
Insurance (No is Reference)
Yes	1.621	1.217 2.159
^*Don’t know	2.186	0.441 10.828
Age category(20 ≤ Age < 30 is Reference)
30 ≤ Age < 40	1.356	1.013 1.816
40 ≤ Age < 50	2.382	1.749 3.245
50 ≤ Age < 60	2.932	2.252 3.816
60 ≤ Age < 70	2.781	2.003 3.861
70 ≤ Age < 80	2.859	2.014 4.059
^*80+	1.375	0.954 1.981
Liver disease (No is Reference)
Yes	1.476	1.060 2.054
^*Don’t know	1.196	0.372 3.847
^*Coronary heart disease (No is Reference)
Yes	1.044	0.824 1.322
^*Kidney disease (No is Reference)
Yes	0.939	0.603 1.461
Do not know	2.136	0.311 14.667

The “refused” category is not shown for the insurance or liver disease results due to low counts.

Not statistically significant.

The odds ratio estimates for the 30- to 50-year age groups show an increasing trend in risk compared with the 20- to 30-year age group. The risk levels for the 50- to 80-year age groups indicate that these participants are nearly three times as likely to have used a prescription NSAID within the previous 30 days compared with those in the youngest category. The odds ratio estimates for subjects aged 80 years and older were not statistically significant.

Discussion

This study aimed to identify potential risks associated with the use of prescription NSAIDs in the U.S. civilian noninstitutionalized population aged 20 years and older. Our analysis revealed specific groups within the population that may be at greater risk from NSAID ADRs. These findings are significant as they can help inform research to develop new approaches for managing pain and inflammation and reduce NSAID-related ADR risks in vulnerable populations. The study results are intended to supplement the FDA’s earlier conclusions and position stated in the 2005 decision memorandum on NSAIDs.

The survey-weighted logistic regression model indicated that women are more likely than men to have used prescription NSAIDs in the past 30 days. This disproportion may be due to the greater prevalence of rheumatoid and osteoarthritis among women.⁹ However, this heightened use of prescription NSAIDs puts women at increased risk of liver injury, gastrointestinal problems, hypertension, and skin rashes. Prescription NSAIDs can be safe for women when used as directed by a health care provider. Still, they may carry certain risks that should be considered, particularly for pregnant or breastfeeding women. Pregnant women should be cautious about taking NSAIDs, particularly during the third trimester of pregnancy, as they can increase the risk of premature closure of a blood vessel in the baby’s heart.²⁹ The FDA recommends pregnant women avoid the use of NSAIDs at 20 weeks and later due to potential problems with low amniotic fluid.³⁰ NSAIDs should also be avoided during labor and delivery, as they can increase the risk of bleeding.

It is worth noting that our estimate for sex and the associated confidence interval is similar to the odds ratio estimate by Davis et al. (OR = 1.603; 95% CI: 1.328, 1.935) in their analysis of older NHANES cycles. This is significant as the older NHANES cycles included participant NSAID OTC use through 2004. NHANES no longer collected NSAID OTC use after 2004 capturing prescription medications exclusively. These similar estimates for sex, along with the work by Yood et al., lend confidence to the production of unbiased estimates in our own study even with available OTC counterparts.

Participants who reported having hypertension or liver disease were, respectively, 82% and 48% more likely to have used prescription NSAIDs than those who did not report these medical conditions, which is concerning since these conditions magnify the harmful side effect risks from the use of these drugs. Individuals with coronary heart or kidney disease were not found to be at greater risk of using prescription NSAIDs compared with those without these conditions. Nonmedical factors such as insurance coverage, age, and race/ethnicity were also associated with prescription NSAID use. Participants who reported having insurance coverage were 62% more likely to have used prescription NSAIDs than those who did not have insurance. Insurance coverage may provide an affordable financial pathway to prescription NSAIDs, especially for those on fixed incomes. Among different racial/ethnic categories, individuals who identified as “other race—multiracial” were found to be more than 56% more likely to use prescription NSAIDs than those who identified as white. While statistically significant, this estimate lacks precision due to the relatively small sample size, as depicted in Supplementary Appendix Table S3. Those who identified as non-Hispanic Asian were 49% less likely to use prescription NSAIDs than those who identified as white. Regarding age, all age categories except for individuals aged 80 years and older were associated with a higher likelihood of prescription NSAID use than the 20- to 30-year age group. The odds ratio for the 50- to 80-year age groups ranged to nearly three times more likely to be users of prescription NSAIDs compared with the youngest group. This may be due to the onset of osteoarthritis and musculoskeletal disorders prevalent in advanced age groups and the need to control pain and inflammation for other aging-related medical issues.³¹

The costs of NSAID-induced ADRs are expected to skyrocket as the U.S. population ages and the prevalence of conditions such as arthritis continues to rise. It is difficult to gauge the total cost for all ADRs due to prescription NSAIDs. Still, specific estimates for the cost of prescription NSAID adverse GI events are projected to exceed four billion dollars per year in the elderly alone.³² Individuals who suffer liver or renal issues related to an NSAID ADR are likely to incur even greater costs.

It is important to note that NHANES is a cross-sectional study, and as such, it cannot establish a temporal relationship between cause and effect. In addition, NHANES does not include data on all ethnic and racial groups. For example, due to the low proportion of American Indians in the population and the challenges associated with collecting data from rural minority groups, information on these subgroups is not available in NHANES. An oversampling design is necessary to obtain data on underrepresented minority groups. As a result, NHANES may not fully represent the diversity in the U.S. population.

It should be noted that NHANES does not at present collect participant usage information on OTC medications. As a result, there is a risk of misclassifying a subject who is exposed to OTC NSAIDs but does not take prescription NSAIDs. However, this type of bias appears to be minimal when the prevalence of prescription NSAID use is low (less than 35%) relative to the overall prevalence of OTC NSAID use.²²

Conclusion

The risks of NSAID ADRs are associated with the dose, duration of use, and sex. Long-term use of prescription NSAIDs carries notable risks that should not be overlooked by clinicians. The results of this study confirm that women and people with medical conditions who would potentially suffer greater harm from NSAID ADRs experience greater exposure to this class of medication. Therefore, it is crucial to continue investigating the safety and efficacy of medications, particularly in specific populations, to reduce the risk of harmful side effects from medication use.

Footnotes

Data Availability Statement

The data supporting this study’s findings are publicly available on the Centers for Disease Control and Prevention website ().

Ethics Statement

The National Center for Toxicological Research Institutional Review Board U.S. Food and Drug Administration approved the study.

Authors’ Contributions

P.R.: Conceptualization (lead) and writing—original draft (lead). B.L.-C.: Conceptualization (supporting). D.W.: Conceptualization (supporting). L.Z.: Writing—review and editing (supporting).

Author Disclosure Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts of interest.

Disclaimer

This article reflects the views of its authors and does not necessarily reflect those of the U.S. Food and Drug Administration. Any mention of commercial products is for clarification only and is not intended as approval, endorsement, or recommendation.

Funding Information

Funding for this project was provided by the National Center for Toxicological Research, Division of Bioinformatics and Biostatistics, U.S. Food and Drug Administration, Jefferson, AR.

Supplementary Material

Supplementary Appendix Table S1

Supplementary Appendix Table S2

Supplementary Appendix Table S3

Supplementary Appendix Table S4

References

Tenenbaum

. The epidemiology of nonsteroidal anti-inflammatory drugs. Can J Gastroenterol, 1999; 13(2):119–122; doi: 10.1155/1999/361651

Watson

, Caster

, Rochon

, et al. Reported adverse drug reactions in women and men: Aggregated evidence from globally collected individual case reports during half a century. EClinicalMedicine, 2019; 17:100188; doi: 10.1016/j.eclinm.2019.10.001

Jenkins

, Seligman

. Decision Memo - Analysis and Recommendations for Agency Action - COX-2 Selective and Non-selective NSAIDs. FDA; Published April 2005. Available from: https://www.fda.gov/drugs/postmarket-drug-safety-information-patients-and-providers/cox-2-selective-includes-bextra-celebrex-and-vioxx-and-non-selective-non-steroidal-anti-inflammatory [Last accessed: April 10, 2023 ].

U. S. Food and Drug Administration. Analysis and recommendations for agency action regarding nonsteroidal antiinflammatory drugs and cardiovascular risk. J Pain Palliat Care Pharmacother, 2005; 19(4):83–97.

Hawboldt

. Adverse events associated with NSAIDs. US Pharmacist, 2008; 33(12):HS5–HS13.

Paulose-Ram

, Hirsch

, Dillon

, et al. Prescription and non-prescription analgesic use among the US adult population: Results from the third National Health and Nutrition Examination Survey (NHANES III). Pharmacoepidemiol Drug Saf, 2003; 12(4):315–326; doi: 10.1002/pds.755

U.S. Food and Drug Administration. Medication guide for non-steroidal anti-inflammatory drugs (NSAIDs). 2007. Available from: https://www.accessdata.fda.gov/drugsatfda_docs/label/2010/018766s015MedGuide.pdf [Last accessed: 17 April, 2023 ].

Neutel

, Maxwell

, Appel

. Differences between males and females in risk of NSAID-related severe gastrointestinal events. Pharmacoepidem Drug Safe, 1999; 8(7):501–507; doi: 10.1002/(SICI)1099-1557(199912)8:7<501::AID-PDS454>3.0.CO;2-Y

Anderson

. Gender differences in pharmacological response. Int Rev Neurobiol, 2008; 83(:1–10; doi: 10.1016/S0074-7742(08)00001-9

10.

Curhan

, Willett

, Rosner

, et al. Frequency of analgesic use and risk of hypertension in younger women. Arch Intern Med, 2002; 162(19):2204–2208; doi: 10.1001/archinte.162.19.2204

11.

Lacroix

, Lapeyre-Mestre

, Bagheri

, et al. General Practitioner Networks. Nonsteroidal anti-inflammatory drug-induced liver injury: A case-control study in primary care. Fundam Clin Pharmacol, 2004; 18(2):201–206; doi: 10.1111/j.1472-8206.2004.00224.x

12.

Tran

, Knowles

, Liu

, et al. Gender differences in adverse drug reactions. J Clin Pharmacol, 1998; 38(11):1003–1009; doi: 10.1177/009127009803801103

13.

Dedier

, Stampfer

, Hankinson

, et al. Nonnarcotic analgesic use and the risk of hypertension in US women. Hypertension, 2002; 40(5):604–608; doi: 10.1161/01.hyp.0000035856.77718.dadiscussion 601-3,

14.

Nakhai-Pour

, Broy

, Sheehy

, et al. Use of nonaspirin nonsteroidal anti-inflammatory drugs during pregnancy and the risk of spontaneous abortion. CMAJ, 2011; 183(15):1713–1720; doi: 10.1503/cmaj.110454

15.

Ostensen

, Skomsvoll

. Anti-inflammatory pharmacotherapy during pregnancy. Expert Opin Pharmacother, 2004; 5(3):571–580; doi: 10.1517/14656566.5.3.571

16.

Demyanets

, Wojta

. Sex differences in effects and use of anti-inflammatory drugs. Handb Exp Pharmacol, 2012(214):443–472; doi: 10.1007/978-3-642-30726-3_20

17.

Dominick

, Ahern

, Gold

, et al. Gender differences in NSAID use among older adults with osteoarthritis. Ann Pharmacother, 2003; 37(11):1566–1571; doi: 10.1345/aph.1C418

18.

Benjeaa

, Geysels

. Gender bias in the clinical evaluation of drugs. Applied Clinical Trials, 2020; 29(12).

19.

Farkouh

, Baumgartel

, Gottardi

, et al. Sex-related differences in drugs with anti-inflammatory properties. J Clin Med, 2021; 10(7); doi: 10.3390/jcm10071441

20.

Davis

, Lee

, Kim

, et al. Use of non-steroidal anti-inflammatory drugs in US adults: Changes over time and by demographic. Open Heart, 2017; 4(1):e000550; doi: 10.1136/openhrt-2016-000550

21.

Chen

, Clark

, Riddles

, et al. National Health and Nutrition Examination Survey, 2015-2018: Sample Design and Estimation Procedures. Vital Health Stat 1, 2020; 184(190):1–36.

22.

Yood

, Campbell

, Rothman

, et al. Using prescription claims data for drugs available over-the-counter (OTC). Pharmacoepidemiol Drug Saf, 2007; 16(9):961–968; doi: 10.1002/pds.1454

23.

Gorczyca

, Manniello

, Pisano

, et al. NSAIDs: Balancing the risks and benefits. US Pharm, 2016; 41(3):24–26.

24.

Austin

. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res, 2011; 46(3):399–424; doi: 10.1080/00273171.2011.568786

25.

DuGoff

, Schuler

, Stuart

. Generalizing observational study results: Applying propensity score methods to complex surveys. Health Serv Res, 2014; 49(1):284–303; doi: 10.1111/1475-6773.12090

26.

Rogers

. Controlling for Confounding in Complex Survey Machine Learning Models to Assess Drug Safety and Risk. In: Machine Learning and Deep Learning in Computational Toxicology. ( Hong

. ed.) Springer International Publishing: Cham; 2023; pp. 355-374.

27.

Faries

, Zhang

, Kadziola

, et al. Real World Health Care Data Analysis: Causal Methods and Implementation Using SAS. SAS Institute Inc: Cary, NC; 2020.

28.

Benedetto

, Head

, Angelini

, et al. Statistical primer: Propensity score matching and its alternatives. Eur J Cardiothorac Surg, 2018; 53(6):1112–1117; doi: 10.1093/ejcts/ezy167

29.

U.S. Food and Drug Administration. FDA has reviewed possible risks of pain medicine use during pregnancy. 2015. Available from: https://www.fda.gov/safety/medical-product-safety-information/nonsteroidal-anti-inflammatory-drugs-nsaids-drug-safety-communication-avoid-use-nsaids-pregnancy-20#:∼:text=These%20labels%20already%20warn%20to,professional%20before%20using%20these%20medicines [Last accessed: 04, April].

30.

U.S. Food and Drug Administration. FDA recommends avoiding use of NSAIDs in pregnancy at 20 weeks or later because they can result in low amniotic fluid. 2020. Available from: https://www.fda.gov/drugs/drug-safety-and-availability/fda-recommends-avoiding-use-nsaids-pregnancy-20-weeks-or-later-because-they-can-result-low-amniotic [Last accessed: 21 August, 2023 ].

31.

Marcum

, Hanlon

. Recognizing the risks of chronic nonsteroidal anti-inflammatory drug use in older adults. Ann Longterm Care, 2010; 18(9):24–27.

32.

Fine

. Quantifying the impact of NSAID-associated adverse events. Am J Manag Care, 2013; 19(14 Suppl):s267–72.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB

Identifying Vulnerabilities to NSAID Adverse Events in the U.S. Population: An Analysis of Preexisting Conditions and Sex

Abstract

Purpose:

Method:

Results:

Conclusion:

Introduction

Methods

Data source

Survey-weighted logistic regression

Weight adjustment for confounding

Results

Descriptive measures

Survey-weighted logistic regression risk estimates

Discussion

Conclusion

Footnotes

Data Availability Statement

Ethics Statement

Authors’ Contributions

Author Disclosure Statement

Disclaimer

Funding Information

Supplementary Material

References

Supplementary Material