Acute Respiratory Infection and Use of Nonsteroidal Anti-Inflammatory Drugs on Risk of Acute Myocardial Infarction: A Nationwide Case-Crossover Study

Abstract

Background.

Previous studies have suggested that acute respiratory infection (ARI) and nonsteroidal anti-inflammatory drugs (NSAIDs) use could trigger acute myocardial infarction (AMI). In some countries, physicians prescribe NSAIDs for patients with ARI for symptom relief. However, there is no research evaluating whether NSAIDs use during ARI episodes may increase the risk of AMI.

Methods.

We identified 9793 patients with an incident hospitalization of AMI (index date) between 2007 and 2011. Using case-crossover design, we compared the following exposure status between the case (1–7-day before index date) and matched control period (366–372-day before index date): NSAIDs use during ARI episodes, ARI episodes without NSAIDs use, NSAIDs use only, or no exposure. Multivariable conditional logistic regression models were used to estimate odds ratios adjusted for potential confounders.

Results.

Nonsteroidal anti-inflammatory drugs use during ARI was associated with a 3.4-fold increased risk of AMI (adjusted odds ratio [aOR] = 3.41; 95% confidence interval [CI] = 2.80–4.16), ARI without NSAIDs use was associated with a 2.7-fold increased risk (aOR = 2.65; 95% CI = 2.29–3.06), and NSAIDs use only was associated with a 1.5-fold increased risk (aOR = 1.47; 95% CI = 1.33–1.62). Moreover, parenteral NSAIDs were associated with much higher risk in ARI patients (aOR = 7.22; 95% CI = 4.07–12.81).

Conclusions.

Nonsteroidal anti-inflammatory drugs use during ARI episodes, especially parenteral NSAIDs, was associated with a further increased risk of AMI.

acute myocardial infarction (AMI), acute respiratory infection (ARI), nonsteroidal anti-inflammatory drugs (NSAIDs), case-crossover study.

(See the editorial commentary by Warren-Gash and Udell, on pages 497–9.)

Recent evidence has suggested that acute respiratory infection (ARI) is associated with an increased risk of acute myocardial infarction (AMI), and the risk of myocardial infarction (MI) is reported to be 2–5-fold greater during the 1st through 7th days of either nonspecified respiratory infection or presumed influenza infection [1–3]. A recent meta-analysis also showed that influenza-like illness (odds ratio [OR] = 2.29; 95% confidence interval [CI] = 1.11–4.73) and respiratory tract infection (OR = 1.89; 95% CI 1.35–2.65) were highly associated with AMI [4].

In clinical practice, nonsteroidal anti-inflammatory drugs (NSAIDs) have been widely used for pain and fever associated with ARI or the common cold in some countries [5]. However, the association of NSAIDs use and the risk of AMI has been reported in many observational studies and clinical trials [6–11]. Existing evidence suggests that not only cyclooxygenase 2 (COX-2) inhibitors but also some nonselective NSAIDs can increase the risk of AMI [12–15]. On 9 July 2015, the US Food and Drug Administration (US FDA) further strengthened its existing label warning that NSAIDs can cause heart attacks, which highlights the need to be alert for cardiovascular adverse events during NSAIDs treatment [16]. In the same year, the Pharmacovigilance Risk Assessment Committee (PRAC) of the European Medicines Agency (EMA) also issued a warning of increased risk of cardiovascular events in patients taking high-dose ibuprofen, with a risk similar to that of COX-2 inhibitors and diclofenac [17].

Because ARI and NSAIDs, individually, are both associated with an increased risk of AMI, it is highly possible that NSAIDs use during ARI episodes, which is common in daily practice, might have a joint effect on the risk of AMI and put patients in a more dangerous situation. However, to the best of our knowledge, no study has ever evaluated this hypothesis. The aim of this case-crossover study was therefore to assess the potential joint effect of ARI and NSAIDs use on the risk of AMI using the National Health Insurance Research Database (NHIRD) in Taiwan.

METHODS

Data Source

The data we used in this study were from the NHIRD, which contains outpatient and inpatient claims, including demographic characteristics, disease diagnoses, and prescription medication records for all beneficiaries enrolled in Taiwan’s mandatory National Health Insurance (NHI) program [18]. All of the claims data for approximately 23 million beneficiaries (approximately 99% of the total population in Taiwan) during the period of 1 January 2005–31 December 2011 were used as the data source. The study protocol was approved by the Research Ethics Committee of National Taiwan University Hospital (NTUH-REC-201406124W).

Study Subjects

Eligible study subjects included patients aged ≥20 years who had an incident hospitalization for AMI, defined by an admission record with a principal diagnosis of AMI [International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM code) 410.xx] between 1 January 2007 and 31 December 2011. The positive predictive value for identifying AMI cases using the principal diagnosis code (ICD-9-CM code 410) has been reported to be 93% in the NHIRD [19]. We defined the index date for each patient as the date of incident hospitalization for AMI. Patients with any outpatient visit or hospitalization for AMI or old myocardial infarction (ICD-9-CM code 412) during the 730 days prior to the index date were excluded to guarantee the enrollment of incident AMI cases in our study. To be eligible for the study, patients also needed to have full NHI coverage for a continuous period of at least 24 months (730 days) prior to the index date to enable the evaluation of patient history. We thus excluded patients without any clinical visit (including outpatient or emergency room visit) during the 730 days prior to the index date. Patients who were admitted for any reason during the 730 days prior to the index date were also excluded because of the uncertainty about the dosage and duration of NSAIDs use during the hospitalization in Taiwan.

Study Design

We used a case-crossover design to investigate the potential joint effect of ARI and use of NSAIDs on the risk of AMI. With the same patients as self-controls, the study could avoid the between-subject differences of measured or unmeasured time-invariant confounders, such as family history or lifestyle factors [20]. We compared the exposure status of individual study subjects between the case period (1–7-day period before the index date) and matched control period (366–372-day period before the index date). To prevent the impact of seasonal effects on the association between ARI and AMI, the control period was defined to be the equivalent time period exactly 1 year (365 days) prior in our study.

Exposure to Acute Respiratory Infection and Nonsteroidal Anti-inflammatory Drugs

An ARI episode was defined as an outpatient visit for ARI (ICD-9-CM code 079.9, 382.9, 460, 461.9, 465.8, 465.9, 466.0, 486, 487.0, 487.1, 487.8, 490, 780.6, or 786.2) during the case or control period. The correlation coefficient between these diagnosis codes and positive respiratory virus pathogens has been reported to be 0.71 [21]. The day of the outpatient visit for an ARI was considered the first day of infection, and for participants with >1 outpatient visit for an ARI in the case or control period, all visits were considered as a single exposure during the period of interest.

Exposure to NSAIDs within the case or control period for the same patient was also assessed to explore the joint effect of ARI and NSAIDs on AMI. The NSAIDs investigated in this study were single-active-ingredient NSAIDs (Anatomical Therapeutic Chemical [ATC] code M01A, except for glucosamine [M01AX05]). Dispensing information of either parenteral or nonparenteral (including oral and rectal) NSAIDs in NHIRD were used in this study. Cumulative exposure to NSAIDs within the case or control period was calculated as defined daily dose, which is defined by the World Health Organization Collaborating Center for Drug Statistics Methodology [22]. The defined daily dose and ATC code of NSAIDs studied are listed in Supplementary Table 1. A high-dose of NSAIDs was defined as a cumulative dose higher than or equal to the 75th percentile of the cumulative dose of our study subjects in the study.

According to the definitions of exposure to ARI and NSAIDs, we categorized the exposure status in either case or control periods as follows: NSAIDs use during ARI episodes, ARI episodes without NSAIDs use, NSAIDs use only, or no exposure.

Baseline Characteristics of Study Subjects

We collected data on age, sex, comorbidities, and concomitant medications during the 1–90-day period and 366–455-day period before the index date for each subject. A detailed list of comorbidities and concomitant medications is presented in Supplementary Table 2.

Statistical Analysis

Multivariable conditional logistic regressions were used to estimate the effect of ARI and NSAIDs use on the risk of AMI. We used the LOGISTIC procedure of SAS software) to calculate the crude odds ratios (cORs) and adjusted odds ratios (aORs) with 95% confidence intervals. The discordant concomitant medications between the case and control periods were considered to be within-subject time-variant covariates for adjusted analysis, including calcium channel blockers, agents acting on the renin-angiotensin system, beta-blockers, diuretics, anti-hypertensive agents, low-dose aspirin, nonaspirin antiplatelet agents, statins, insulin, and acetaminophen. A preplanned secondary analysis was conducted to investigate the effects of routes of administration (ie, parenteral or nonparenteral) and cumulative dose of NSAIDs on the risk of AMI.

Sensitivity analyses were performed to examine the robustness of the results by varying the time window of exposure from 7 days (case = days 1–7; control = days 366–372) to 14 days (case = days 1–14; controls = days 366–379) and 30 days (case = days 1–30; controls = days 366–395) without changing the definition of the index date.

RESULTS

We identified 9793 patients with an incident admission for AMI between 2007 and 2011 who fulfilled all eligibility criteria (Figure 1). Among the patients, 61.35% were male, and the mean age at the index date was 72.29 years old (standard deviation [SD] = 12.65). Their comorbidities and concomitant medications in the case and control periods are summarized in Table 1. Diabetes mellitus was the most frequent comorbidity (42.59% in the case period and 40.71% in the control period), followed by hypertension (41.68% in the case period and 39.93% in the control period). Calcium channel blocker was the most frequent comedication (50.07% in the case period and 47.84% in the control period), followed by agents acting on the renin-angiotensin system (49.88% in the case period and 46.66% in the control period). Compared with the control period, a higher proportion of subjects had comorbidities and comedications in the case period.

Figure 1.

Study flowchart.

Table 1.

Comorbidities and Concomitant Medications During the 1–90-Day Period and 366–455-Day Period Before the Index Date (N = 9793)

	1–90-day period before the index date	366–455-day period before the index date	P value
Comorbidity, no. (%)
Cerebrovascular disease	1677 (17.12)	1550 (15.83)	<.0001
Hypertension	4082 (41.68)	3910 (39.93)	.0001
Congestive heart failure	1726 (17.62)	1232 (12.58)	<.0001
Valvular heart disease	348 (3.55)	295 (3.01)	.0005
Atrial fibrillation	355 (3.63)	261 (2.67)	<.0001
Peripheral vascular disease	324 (3.31)	259 (2.64)	.0002
Diabetes mellitus	4171 (42.59)	3987 (40.71)	<.0001
Lipid disorder	1856 (18.95)	1884 (19.24)	.41
Peptic ulcer disease	1226 (12.52)	1033 (10.55)	<.0001
Chronic pulmonary disease	1657 (16.92)	1236 (12.62)	<.0001
Chronic renal disease	1911 (19.51)	1504 (15.36)	<.0001
Chronic liver disease	298 (3.04)	259 (2.64)	.03
Osteoarthritis	1232 (12.58)	1194 (12.19)	.30
Rheumatoid arthritis	80 (0.82)	89 (0.91)	.22
Migraine	26 (0.27)	31 (0.32)	.43
Cancer	618 (6.31)	449 (4.58)	<.0001
Concomitant medication, no. (%)
Calcium channel blockers	4903 (50.07)	4685 (47.84)	<.0001
Agents acting on the RAS	4885 (49.88)	4569 (46.66)	<.0001
Beta-blockers	3797 (38.77)	3386 (34.58)	<.0001
Diuretics	4304 (43.95)	3753 (38.32)	<.0001
Antihypertensive agents	1196 (12.21)	1125 (11.49)	.02
Low-dose aspirin	4461 (45.55)	3752 (38.31)	<.0001
Non-aspirin antiplatelet agents	3153 (32.20)	2199 (22.45)	<.0001
Vitamin K antagonists	287 (2.93)	273 (2.79)	.29
Statins	2541 (25.95)	2415 (24.66)	.001
Insulin	1711 (17.47)	1336 (13.64)	<.0001
Oral hypoglycemic agents	3546 (36.21)	3560 (36.35)	.63
Acetaminophen	4438 (45.32)	3685 (37.63)	<.0001

Abbreviation: RAS, renin-angiotensin system

Table 2 shows that NSAIDs use during ARI episodes was associated with a 3.4-fold increased risk of AMI (aOR = 3.41; 95% CI = 2.80–4.16) after adjusting for discordant concomitant medications between the case period and control period. ARI episodes without NSAIDs use or NSAIDs use only were associated with a 2.7-fold (aOR = 2.65; 95% CI = 2.29–3.06) or 1.5-fold (aOR = 1.47; 95% CI = 1.33–1.62) increased risk of AMI, respectively.

Table 2.

Risk of Acute Myocardial Infarction Associated With Acute Respiratory Infection and/or Nonsteroidal Anti-inflammatory Drugs Use (N = 9793)

Exposure	Exposed during the case period, no.	Exposed during the control period, no.	Crude odds ratio	95% CI	Adjusted odds ratioa	95% CI
NSAIDs use during ARI episodes	470	176	3.77	3.11–4.56	3.41	2.80–4.16
ARI episodes without NSAIDs use	767	334	2.85	2.48–3.28	2.65	2.29–3.06
NSAIDs use only	1516	1235	1.58	1.44–1.74	1.47	1.33–1.62
No exposure	7040	8048	1.00	…	1.00	…

Abbreviations: ARI, acute respiratory infection; CI, confidence interval; NSAIDs, nonsteroidal anti-inflammatory drugs.

^aAdjusted for discordant use of concomitant medications between the 1–90-day period and 366–455-day period before the index date.

Parenteral NSAIDs use during ARI episodes was associated with the highest risk of AMI (aOR = 7.22; 95% CI = 4.07–12.81). High-dose nonparenteral NSAIDs use during ARI episodes was associated with a 3.3-fold increased risk of AMI (aOR = 3.32; 95% CI = 2.34–4.93). Even low-dose nonparenteral NSAIDs use during ARI episodes was associated with an approximately 3-fold increased risk of AMI (aOR = 2.95; 95% CI = 2.31–3.75) (Table 3). Sensitivity analyses by varying the length of the exposure time window yielded similar results (Table 4).

Table 3.

Risk of Acute Myocardial Infarction Associated With Acute Respiratory Infection and/or Nonsteroidal Anti-inflammatory Drugs (NSAIDs) Use by Routes of Administration and Cumulative Dose of NSAIDs (N = 9793)

Exposure			Exposed during the case period, no.	Exposed during the control period, no.	Adjusted odds ratioa	95% CI
ARI episodes	Parenteral NSAIDs		94	15	7.22	4.07–12.81
	Nonparenteral	High-doseb NSAIDs	108	48	3.32	2.34–4.93
		Low-dosec NSAIDs	268	113	2.95	2.31–3.75
	No NSAIDs		767	334	2.65	2.29–3.07
No ARI	Parenteral NSAIDs		249	81	3.77	2.83–5.02
	Nonparenteral	High-doseb NSAIDs	298	328	1.10	.92–1.32
		Low-dosec NSAIDs	969	826	1.38	1.23–1.54
	No NSAIDs		7040	8048	1.00	…

Abbreviations: ARI, acute respiratory infection; CI, confidence interval; NSAIDs, nonsteroidal anti-inflammatory drugs.

^aAdjusted for discordant use of concomitant medications between the 1–90-day period and 366–455-day period before the index date.

^bHigh-dose: Cumulative dose within time window ≥8 defined daily dose.

^cLow-dose: Cumulative dose within time window >0 defined daily dose and <8 defined daily dose.

Table 4.

Sensitivity Analyses on Different Length of Time Windows (N = 9793)

Exposure	Exposed during the case period, no.	Exposed during the control period, no.	Crude odds ratio	95% CI	Adjusted odds ratioa	95% CI
Time window = 7 days
NSAIDs use during ARI episodes	470	176	3.77	3.11–4.56	3.41	2.80–4.16
ARI episodes without NSAIDs use	767	334	2.85	2.48–3.28	2.65	2.29–3.06
NSAIDs use only	1516	1235	1.58	1.44–1.74	1.47	1.33–1.62
No exposure	7040	8048	1.00	...	1.00	...
Time window = 14 days
NSAIDs use during ARI episodes	688	343	2.88	2.48–3.35	2.59	2.22–3.03
ARI episodes without NSAIDs use	959	527	2.37	2.10–2.67	2.16	1.90–2.46
NSAIDs use only	1667	1441	1.51	1.38–1.65	1.39	1.27–1.53
No exposure	6479	7482	1.00	…	1.00	…
Time window = 30 days
NSAIDs use during ARI episodes	1099	659	2.43	2.15–2.74	2.15	1.89–2.44
ARI episodes without NSAIDs use	1169	798	1.91	1.71–2.12	1.73	1.54–1.93
NSAIDs use only	1858	1775	1.36	1.24–1.48	1.25	1.14–1.37
No exposure	5667	6561	1.00	…	1.00	…

Abbreviations: ARI, acute respiratory infection; CI, confidence interval; NSAIDs, nonsteroidal anti-inflammatory drugs.

^aAdjusted for discordant use of concomitant medications between the 1–90-day period and 366–455-day period before the index date.

DISCUSSION

Our findings support our hypothesis that having 2 risk factors for AMI (ie, ARI and NSAIDs) would further increase the AMI risk compared with having only 1 risk factor for AMI (ie, ARI only or NSAIDs only). To the best of our knowledge, this study is the first to investigate the joint effect of ARI and NSAIDs use on the risk of AMI. This approach should raise clinical concern because NSAIDs use during ARI episodes is highly common in real-world practice. For example, a previous study showed that the most frequent diagnosis for NSAID prescriptions is acute nasopharyngitis (ICD-9-CM code 460) [23]. Instead of treating respiratory infection as a confounding factor as in some previous studies that focused on the association between NSAIDs use and the risk of cardiovascular disease [17, 18], our study identified ARI as an individual exposure to reveal the joint effect of ARI and NSAIDs use on the risk of AMI.

The mechanism behind the potential joint effect of ARI and NSAIDs could be explained by existing evidence. Several studies found that infection induces the biosynthesis of proinflammatory and prothrombotic cytokines and increases the accumulation of macrophages in atherosclerotic lesions [24, 25]. In addition, ARI induces systemic inflammation and coagulation, leading to thrombosis [26, 27]. Although NSAIDs are thought to have an anti-inflammatory effect, inhibition of cyclooxygenase would decrease the level of antithrombotic prostaglandin I₂ and enhance the synthesis of leukotrienes, leading to platelet aggregation and vasoconstriction. Furthermore, NSAIDs can cause elevation of blood pressure by sodium and water retention and lead to plaque rupture [28, 29]. Therefore, NSAIDs use during an ARI would increase the occurrence of AMI mechanistically.

Our findings concerning ARI only or NSAIDs only and the risk of AMI are similar to those of previous studies. Previously, database research in the United Kingdom using case-crossover analysis showed a relative risk of 2.7 (95% CI = 1.6–4.7) for AMI related to ARI in the 10 days before the index date [30], which was quite similar to the risk associated with ARI only (aOR = 2.65; 95% CI = 2.29–3.06) in our study, although the length of time window was somewhat different. The risk associated with NSAIDs only in our study (aOR = 1.47; 95% CI = 1.33–1.62) was also similar to that of a previous case-crossover study in Taiwan, which found that oral NSAIDs were associated with a 1.42-fold increased risk of AMI (95% CI = 1.29–1.56) [23]. These similarities makes it more convincing that NSAIDs use during ARI episodes would further increase the risk of AMI.

Statistically, there was a negative multiplicative interaction between ARI and NSAIDs (the 2 risk factors of AMI) because the observed joint adjusted odds ratio was less than the expected value based on independent adjusted odds ratios [31, 32]. A possible explanation was that NSAIDs have an anti-inflammatory effect that attenuates the increased risk of AMI associated with ARI. However, the magnitude of the risk of AMI associated with NSAIDs use during ARI episodes was still higher than the independent effects of ARI or NSAIDs on the risk of AMI, indicating that NSAIDs might trigger AMI through a different mechanism.

Our study also suggested that the risk of AMI in patients using parenteral NSAIDs was higher than that observed in patients using nonparenteral NSAIDs, which is consistent with previous studies [23, 33]. For example, a study that also had a case-crossover study design found that the risk of AMI was 3.4-fold higher in patients using parenteral NSAIDs (aOR = 3.35; 95% CI = 2.50–4.47) [23]. Furthermore, we found that the risk of AMI associated with parenteral NSAIDs use during ARI episodes was even higher. Although ibuprofen was the only parenteral NSAIDs approved for fever relief by the US FDA, there were several other NSAIDs that are available in parenteral form, such as ketorolac and diclofenac. Therefore, the safety issue of parenteral NSAIDs use during ARI is still noteworthy.

For nonparenteral NSAIDs, we observed that during ARI episodes, a high dose of NSAIDs was associated with a higher risk of AMI than a low-dose of NSAIDs. This result was consistent with previous studies showing that high-dose NSAIDs use was associated with a greater risk of AMI than low-dose use [23, 33]. In our study, however, without considering ARI, high-dose NSAIDs use was not highly associated with a higher risk of AMI than low-dose NSAIDs. A possible explanation for this is that many patients without ARI were prescribed high-dose NSAIDs for their underlying diseases, so the exposure status to NSAIDs was almost the same in the case and control period, resulting in an insignificant association between high-dose NSAIDs use and AMI. We examined our data and found that approximately 40% of those who received high-dose NSAIDs without ARI had osteoarthritis (38.26% in the case period and 41.77% in the control period), which supports our inference.

Similar to other observational studies based on claims databases, our study has several limitations. First, some information about the study subjects is not assessable from the database. For example, it is not easy to perfectly capture the severity of ARI in our study because laboratory data are not recorded in the claims data we used. However, we have conducted a sensitivity analysis using ARI types as a surrogate of ARI severity to estimate the potential confounding by indication. We found that NSAIDs use during influenza-related ARI episodes was associated with the highest risk of AMI (Supplementary Table 3). More epidemiology studies are needed to investigate the influence of infection severity on cardiovascular risk. In addition, we were unable to ascertain that dispensed NSAIDs were actually taken by the participants in a claims-based study. There might be nondifferential misclassification bias in our study because the NHIRD does not have information concerning nonprescription drugs use, which might underestimate the association between NSAIDs and the risk of AMI.

Second, although we adopted a case-crossover design and adjusted for discordant concomitant medications, we were not able to control for all confounding factors. However, we yielded similar results after further adjusting for discordant comorbidities, including cerebrovascular disease, hypertension, congestive heart failure, valvular heart disease, atrial fibrillation, peripheral vascular disease, diabetes mellitus, peptic ulcer disease, chronic pulmonary disease, chronic renal disease, chronic liver disease, and cancer (data not shown). Third, our study might have protopathic bias because physicians might prescribe NSAIDs for certain symptoms before the occurrence of AMI. However, our sensitivity analyses by varying the length of the exposure time window yielded similar results. We believe that the extent of protopathic bias would be minor. Lastly, we did not present results regarding the potential different effects of different NSAIDs use during ARI on risk of AMI. Nevertheless, we have conducted an exploratory analysis to address this issue (Supplementary Table 4). We did observe different effects of different NSAIDs during ARI on risk of AMI. However, these findings should be interpreted with caution due to the relatively small sample size and wide confidence interval.

In conclusion, we determined that NSAIDs use during ARI episodes, particularly parenteral NSAIDs use, was associated with a further increased risk of AMI.

Notes

Acknowledgments. We thank the National Health Insurance Administration (NHIA) and National Health Research Institutes (NHRI) for making available the databases used in this study. W. Y. C., H. F. Y., S. L. J., and F. C. C. contributed to the study concept and design. W. Y. C., L. Z. F., and H. F. Y. contributed to the acquisition of data and statistical analysis. H. F. Y., S. L. J., and F. C. C. contributed to the interpretation of data. W. Y. C. and H. F. Y. contributed to drafting the manuscript. H. F. Y., C. K. A., S. L. J., and F. C. C. contributed to revision of the manuscript. All authors read and approved the final manuscript.

Disclaimer. The content of this article does not represent any official position of the NHIA or NHRI. The author had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Financial support. This work was supported by the Ministry of Science and Technology, Taiwan (Grant No. MOST103-2320-B-002-015) and National Taiwan University Hospital (Grant No. NTUH105-A129). There is no relationship with industry.

Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Author notes

a

Y.-C. W. and F.-Y. H. contributed equally to the study.

Correspondence: C.-C. Fang, MD, Department of Emergency Medicine, National Taiwan University Hospital, Taipei, Taiwan, No. 7, Chung-Shan South Rd, Taipei 100, Taiwan ([email protected]).