Abstract
Background
Inappropriate antibiotic use in acute respiratory infections (ARIs) is a major public health concern; however, data for people with human immunodeficiency virus (PWH) are limited.
Methods
The HIV Virtual Cohort Study is a retrospective cohort of adult Department of Defense beneficiaries. Male PWH cases (n = 2413) were matched 1:2 to controls without HIV (n = 4826) by age, gender, race/ethnicity, and beneficiary status. Acute respiratory infection encounters between 2016 and 2020 and corresponding antibiotic prescriptions were characterized as always, sometimes, or never appropriate based on International Classification of Diseases, Tenth Revision coding. Incidence of ARI encounters and antibiotic appropriateness were compared between PWH and controls. Subgroup analyses were assessed by CD4 count and viral load suppression on antiretroviral therapy.
Results
Mean rates of ARI encounters were similar for PWH (1066 per 1000 person-years) and controls (1010 per 1000 person-years); however, the rate was double among PWH without viral load (VL) suppression (2018 per 1000 person-years). Antibiotics were prescribed in 26% of encounters among PWH compared to 34% for controls (P ≤ .01); antibiotic use was “never” appropriate in 38% of encounters with PWH and 36% in controls. Compared to controls, PWH received more sulfonamides (5.5% vs 2.7%; P = .001), and variation existed among HIV subgroups in the prescription of sulfonamides, fluoroquinolones, and β-lactams.
Discussion
Acute respiratory infection encounters were similar for PWH and those without HIV; however, PWH with lower CD4 counts and/or nonsuppressed VL had more frequent ARI visits. Inappropriate antibiotic use for ARIs was high in both populations, and focused interventions to improve antibiotic appropriateness for prescribers caring for PWH should be pursued.
Keywords: antibiotics, HIV, outpatient, upper respiratory infection
When comparing antibiotic prescriptions for acute respiratory infections in those with HIV and those without HIV, differences in antibiotic prescription rates were seen. Many prescriptions were universally inappropriate, but differences in prescribed antibiotic classes were noted.
Antibiotic resistance is an international health emergency that is driven by the inappropriate prescribing of antibiotics [1]. Addressing this issue features prominently in multiple national and international public health strategies [2–4]. Previous studies in outpatient and acute care populations have shown antibiotic prescriptions are widely variable with as many as 50% being inappropriate prescriptions [5–12]. Acute respiratory infections (ARIs) in adults are most likely to result in inappropriate prescribing and thus have the greatest value as a target for improvement in the reduction of inappropriate prescriptions [5].
In prior studies, researchers have attempted to quantify the prescribing patterns surrounding antibiotics in a variety of clinical situations. Substantial variability in prescribing rates has been noted, with one study of all outpatient prescribing for all indications reporting as high as 1115 prescriptions per 1000 person-years (p-y) [13], whereas another study evaluating all antibiotics in the primary care setting reported a more modest 84 per 1000 p-y [8]. In terms of prescribing rates for antibiotics during outpatient ARI encounters, these generally range from 30% to 90% of visits in different settings [9, 14–17]. The appropriateness of these antibiotic prescriptions also varies considerably. For example, multiple Veterans Administration studies demonstrated that inappropriate antibiotic prescriptions for ARIs appears to be as high 76% and frequently not lower than the 60th percentile [8, 9, 15]. This discrepancy in inappropriate prescribing patterns may be partly due to the use of antibiotics that are either guideline discordant in duration or class of antibiotic [8]. As such, fluoroquinolones and macrolides were the most frequently overprescribed or inappropriately prescribed antibiotics, each accounting for up to 20% of prescriptions in many studies [8, 13, 16, 18, 19].
Despite the recent focus on this issue and a clear need for improving guideline adherence, our knowledge of prescribing patterns among subpopulations remains low and mostly unaddressed. Patient comorbidities have previously been shown to be a significant influencer of antibiotic prescribing practices [20], yet many patients with comorbidities are often excluded from analyses of antibiotic prescriptions. One such high-risk population is people with human immunodeficiency virus (PWH). Due to concerns about immunosuppression and the common use of prophylactic antibiotics, PWH are often excluded from these analyses of antibiotic prescriptions, and only sparse, limited population data specifically addressing use in PWH exists [21–23]. The existing data are mixed, sometimes showing PWH receiving similar frequency of antibiotics to a non-HIV-infected population [21], and sometimes significantly more [22, 23]. Furthermore, whether patterns of prescribing in this population vary by stage of HIV disease, type of ARI, or class of antibiotic has also not been systematically studied among a large, matched cohort. To address these knowledge gaps, we sought to evaluate the incidence of ARIs and antibiotic prescription patterns in PWH compared to those without HIV infection in the Department of Defense (DoD) HIV Virtual Cohort Study (VCS).
METHODS
A retrospective matched virtual cohort analysis was performed through the selection of participants from the VCS, which comprises DoD beneficiaries with and without HIV infection. Inclusion criteria in the cohort include age 18 and older, data within the military health system available for at least 1 year (defined as at least 2 encounters at least 1 year apart between 1997 and 2020). Cases for this analysis were defined as adult male PWH (due to small number of female HIV + service members and retirees) and presence/absence of HIV diagnosis based on International Classification of Diseases, Nineth Revision (ICD-9) and International Classification of Diseases, Tenth Revision (ICD-10) codes. Controls were defined as adult males without HIV infection. People with HIV were oversampled to include all such individuals, and the control subset was randomly sampled in a stratified manner to be matched to the PWH population in a 2:1 ratio based on beneficiary status, race, age at first encounter, and sex. Since the cohort is predominantly male, women were excluded from this analysis.
The period of the fiscal year 2016–2020 was selected for this study to correspond with the most recent Infectious Diseases Society of America guidelines on Antibiotic Stewardship [24] as well as exclusive use of ICD-10 coding in the medical record. The ICD-10 codes for all outpatient clinical encounters for ARI diagnoses were selected and included as previously described [18]. In brief, all ICD-10 codes associated with ARIs were used. The appropriateness of antibiotic usage was determined per protocols designed by Fleming-Dutra et al [5] and Chua et al [18], in which oral outpatient antibiotic usage for each ICD code was given a classification as “always”, “sometimes”, or “never” appropriate based on the most up-to-date clinical guidelines, followed by expert consensus among multiple internal medicine and pediatric physicians when such guidelines were otherwise lacking. Per the previously described schema, the assigning of antibiotic appropriateness erred on the side of assuming appropriateness. For example, diagnosis codes were considered as “sometimes” appropriate for antibiotics when they denoted nonspecific categories that could refer to either viral or bacterial etiologies (in which a viral etiology suggest an inappropriate use of antibiotics) [18]. If multiple ICD-10 codes were associated with 1 encounter, then a classification of “never” was used only when codes for “always” or “sometimes” were not present. Analyses were dichotomized as “always/sometimes” versus “never” appropriate to account for instances in which multiple ICD-10 with different levels of suitability were present. Demographic, prescription patterns, clinical diagnoses, and HIV-related characteristics were analyzed. Smoking status was also broadly captured using a binomial “never/ever smoker” signifier because tobacco usage in the cohort is variably reported. If any cigarette/inhaled-tobacco usage was reported at any point in the studied period, the participant was given an “ever smoker” designation.
Among PWH, subgroup analyses were conducted using subcategories at the time of each encounter including CD4 count (≥500 cells/µL vs <500 cells/µL), presence of viral load (VL) suppression (defined as VL <200 copies/mL) on antiretroviral therapy (ART), and by “optimized” treatment status (defined as VL <200 copies/mL on ART with CD4 ≥500 cells/µL) versus “nonoptimized” (defined as VL >200 copies/mL, not on ART, or with CD4 <500 cells/µL). A CD4 count of >500 was chosen, because prior research indicates that these individuals may have life-spans approaching uninfected individuals [25]. In addition, the cohort of those with CD <200 did not provide a large enough pool to power analyses. If no laboratory values were present at the time of the ARI encounter, the most recent preceding laboratory sets were used to better match the clinically available data at the time of encounter. Stata (StataCorp 2019, Stata Statistical Software, Release 16; StataCorp LLC, College Station, TX) was used to compare groups using Student t test or χ2 as appropriate.
Patient Consent Statement
This study was approved by the Uniformed Services University of the Health Sciences Institutional Review Board (IRB) (FWA 00001628; DoD Assurance P60001) Approval of Protocol IDCRP-091 for Human Subjects Participation. The research has been approved by Uniformed Services University of the Health Sciences and National Institutes of Health IRBs. Patient consent was not obtained due to the retrospective design of the study. All data were deidentified for analysis.
RESULTS
A total of 2413 PWH and 4827 controls were included. Most subjects were nonactive-duty retirees or beneficiaries (66.4%) with a mean age of 45.9 (±12.8) years and were mostly White (44.0%) or African American (43.1%) (Table 1). Among PWH, the mean CD4 count at ARI encounters was 651 (±331) cells/µL with 87.4% on ART. History of tobacco use was higher for PWH (45.1%) compared to controls (29.9%; P < .001). The total number of encounters was 4962 for PWH and 6225 for controls. The top 4 most recorded ARI-related ICD-10 codes were similar for both groups and accounted for 49.3% of codes for PWH and 50.7% of codes for the control group. Acute respiratory infection (J06.9) was the most utilized ICD-10 code for both groups representing 16.8% of all ICD-10 codes used. Among both controls and PWH, the top 5 codes representing 50% and 53% of all coded ICD-10 codes were the same (acute respiratory infection [J06.9], pneumonia of unspecified organism [J18.9], acute pharyngitis [J02.9], acute sinusitis [J01.9], and acute bronchitis, unspecified [J20.9]). The top 80% of ICD-10 codes are showing in Table 2.
Table 1.
Characteristics of VCS Participants
| Characteristics | PWH | Controls | P Value |
|---|---|---|---|
| Age at first encounter | 46.2 | 45.6 | .18 |
| Gender, male | 2413 (100) | 4827 (100) | |
| Race/ethnicity | .707 | ||
| White | 1056 (43.8) | 2106 (43.6) | |
| African American | 1051 (43.6) | 2075 (43.0) | |
| Other | 306 (12.7) | 645 (13.4) | |
| Military Status | 1.000 | ||
| Active duty, officer | 120 (5.0) | 240 (5.0) | |
| Active duty, enlisted | 692 (28.7) | 1384 (28.7) | |
| Nonactive duty beneficiary | 1601 (66.4) | 3202 (66.4) | |
| History of ever tobacco use | 1087 (45.1) | 1445 (29.9) | .000 |
| HIV characteristics | … | ||
| Mean age at HIV diagnosis (years) | 36.2 (±10.3) | N/A | |
| Mean time from last negative to first positive HIV test (years) | 10 (±7.2) | N/A | |
| Mean CD4 count at HIV diagnosis (cells/µL) | 679.3 (±316.5) | N/A | |
| Mean CD4 count at first encounter (cells/µL) | 651.8 (±330.5) | N/A | |
| CD4 count ≥500 cells/µL | 1865 (77.29%) | N/A | |
| Mean VL at HIV diagnosis (log10 copies/mL) | −2.1 (±3.6) | N/A | |
| Mean time from HIV diagnosis to ART (years) | 2.5 (±3.3) | N/A | |
| On ART at/before first encounter | 2090 (87.4) | N/A | |
| Suppressed VL (<200 copies/mL) at first encounter | 2054 (85.1) | N/A |
Abbreviations: ART, antiretroviral therapy; HIV, human immunodeficiency virus; N/A, not applicable; PWH, people with HIV; VL, viral load.
Note: All data expressed as number (%) or mean (±standard deviation).
Table 2.
Top ICD-10 Codes
| Controls | PWH | |||||
|---|---|---|---|---|---|---|
| ICD-10 Code | Instances | % of total | ICD-10 Code | Instances | % of total | |
| J06.9 | 1113 | 16.75 | J06.9 | 1018 | 16.82 | |
| J02.9 | 680 | 10.23 | J18.9 | 879 | 14.52 | |
| J01.90 | 526 | 7.92 | J02.9 | 715 | 11.81 | |
| J20.9 | 513 | 7.72 | J01.90 | 334 | 5.52 | |
| J18.9 | 495 | 7.45 | J20.9 | 320 | 5.29 | |
| J01.00 | 322 | 4.85 | J00 | 233 | 3.85 | |
| J32.9 | 310 | 4.67 | J32.9 | 190 | 3.14 | |
| J32.9 | 310 | 4.67 | J31.0 | 185 | 3.06 | |
| J00 | 263 | 3.96 | J02.0 | 177 | 2.92 | |
| J31.0 | 216 | 3.25 | J01.00 | 148 | 2.45 | |
| J40 | 207 | 3.12 | J11.1 | 130 | 2.15 | |
| J02.0 | 142 | 2.14 | J11.1 | 130 | 2.15 | |
| J11.1 | 136 | 2.05 | J18.1 | 124 | 2.05 | |
| J32.0 | 109 | 1.64 | J40 | 118 | 1.95 | |
| … | … | … | J10.1 | 87 | 1.44 | |
| … | … | … | J32.0 | 79 | 1.31 | |
| … | … | |||||
| Total | 6644 | 80.40% | … | 6053 | 80.40% | |
Abbreviations: ICD, International Classification of Diseases, Tenth Edition; PWH, people with human immunodeficiency virus.
Note: Multiple ICD-10 codes may be entered during a single visit. “Total” indicates all ICD-10 codes included in the analysis.
The overall rate of ARI encounters for PWH and controls was similar between groups with 1066 and 1010 visits per 1000 p-y (P = .821), respectively (Table 3). Among PWH, individuals with nonsuppressed VL had higher mean ARI encounters (2018 encounters/1000 p-y) than those with suppressed VL (865 encounters/1000 p-y, P < .001). People with HIV with CD4 count <500 cells/µL also had greater mean ARI encounters with 1878 encounters/1000 p-y versus 847 encounters/1000 p-y (P < .001), respectively. Higher mean encounters were also observed for nonoptimized (1532 encounters/1000 p-y) compared to those who were medically optimized (836 encounters/1000 p-y; P = .005) (Table 3).
Table 3.
Mean ARI Encounters by HIV Disease Characteristics
| Characteristics | PWH | Controls | P Value |
|---|---|---|---|
| Mean ARI encounters/1000 person-years | |||
| All subjects | 1066 (±4285) | 1010 (±8379) | .821 |
| VL nonsuppressed | 2018 (±9929) | — | <.01 |
| VL suppressed | 865 (±1096) | — | — |
| CD4 <500 cells/µL | 1878 (±9049) | — | <.01 |
| CD4 ≥500 cells/µL | 847 (±1021) | — | — |
| Nonoptimized | 1532 (±7279) | — | <.01 |
| Optimized | 836 (±1073) | — | — |
Abbreviations: ARI, acute respiratory infection; ART, antiretroviral therapy; PWH, people with human immunodeficiency virus; VL, viral load.
Note: All data expressed as mean (±standard deviation).
For all ARI encounters, the percentage of encounters at which antibiotics were prescribed for PWH (1752 encounters with prescriptions of 4962 total encounters, 26.1% of encounters) was lower compared to controls (2451 encounters with prescriptions of 6225 total encounters, 34.2% of encounters) and was statistically significant (P ≤ .01). Analyses of antibiotic appropriateness revealed that a similar high proportion of antibiotic prescribing for present ICD diagnoses was “never” appropriate for both PWH (38.0%) and controls (35.5%; P = .09), and although PWH had slightly more inappropriate prescriptions, it was not significant (Table 4). Antibiotic appropriateness did not differ among PWH based on CD4 count, viral load, or optimized status.
Table 4.
Antibiotic Appropriateness for PWH Compared to Controls
| Group/Subgroup | All Encounters | ||
|---|---|---|---|
| Always/Sometimes | Never | P Value | |
| All Antibiotics | .09 | ||
| All PWH | 1079 (61.6) | 666 (38.0) | |
| All Controls | 1575 (64.3) | 871 (35.5) | |
| PWH Subgroups | |||
| Viral load | … | … | .65 |
| VL Nonsuppressed | 217 (62.9) | 128 (37.1) | |
| VL suppressed | 862 (61.6) | 538 (38.4) | |
| CD4 count | … | … | .11 |
| CD4 <500 cells/µL | 248 (58.6) | 175 (41.4) | |
| CD4 ≥500 cells/µL | 831 (62.9) | 491 (37.1) | |
| Optimized treatment Statusa | … | … | .32 |
| Nonoptimized | 382 (60.3) | 251 (39.7) | |
| Optimized | 697 (62.7) | 415 (37.3) | |
Abbreviations: ARI, acute respiratory infection; ART, antiretroviral therapy; PWH, people with human immunodeficiency virus; VL, viral load.
Optimized treatment status defined as VL <200 copies/mL on ART with CD4 ≥500 cells/µL. All data expressed as number (%).
Compared to controls, PWH received more prescriptions for sulfonamides (8.4% vs 6.4%, P = .01) but otherwise had similar antibiotic class profiles (Table 5). Most differences existed within the HIV-subgroup analysis. Those with nonsuppressed VLs were more likely to have prescriptions for β-lactams (19.7% vs 13.7%, P ≤ .01) and clindamycin (4.9% vs 1.6%, P≤.01) but less fluoroquinolones (11.6% vs 18.3%, P ≤ .01). Those with CD4 counts of <500 cells/µL were more likely to have fluoroquinolones (20.4% vs 15.7%, P = .02) and sulfonamides (15.3% vs 6.2%, P ≤ .01) prescribed. Those considered “nonoptimized” were more likely to have sulfonamides (11.3% vs 6.6%, P ≤ .01) and less likely to have β-lactam (8.5% vs 12.4%, P = .01) antibiotics.
Table 5.
Antibiotic Prescriptions by Class
| Antibiotic Class/Antiobiotic | PWH | Control | P Value | VL Nonsuppressed | VL Suppressed | P Value | CD4 <500 Cells/µL | CD4 ≥500 Cells/µL | P Value | Nonoptimized | Optimized | P Value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All Encounters | ||||||||||||
| Macrolide | 598 (34.0) | 861 (35.1) | .46 | 121 (35.1) | 477 (33.9) | .67 | 146 (34.4) | 452 (34.1) | .91 | 207 (32.6) | 391 (35.0) | .31 |
| β-lactam/β-lactamase inhibitor | 198 (11.3) | 266 (10.9) | .68 | 29 (8.4) | 163 (11.6) | .09 | 32 (7.5) | 160 (12.1) | <.01 | 54 (8.5) | 138 (12.4) | .01 |
| Other β-lactam | 261 (14.9) | 412 (16.8) | .1 | 68 (19.7) | 193 (13.7) | <.01 | 44 (10.4) | 217 (16.4) | <.01 | 100 (15.7) | 161 (14.4) | .46 |
| Fluroquinolone | 297 (16.9) | 407 (16.6) | .8 | 40 (11.6) | 257 (18.3) | <.01 | 88 (20.7) | 209 (15.7) | .02 | 115 (18.1) | 182 (16.3) | .33 |
| Tetracycline | 217 (12.3) | 312 (12.7) | .7 | 38 (11.0) | 179 (12.7) | .39 | 48 (11.3) | 169 (12.7) | .45 | 69 (10.9) | 148 (13.2) | .16 |
| Sulfonamide | 147 (8.4) | 157 (6.4) | .01 | 32 (9.3) | 115 (8.2) | .51 | 65 (15.3) | 82 (6.2) | <.01 | 73 (11.5) | 74 (6.6) | <.01 |
| Clindamycin | 40 (2.3) | 36 (1.5) | .46 | 17 (4.9) | 23 (1.6) | <.01 | 2 (0.5) | 38 (2.9) | <.01 | 17 (2.7) | 23 (2.1) | .42 |
Abbreviations: ART, antiretroviral therapy; PWH, people with human immunodeficiency virus; VL, viral load.
Notes: All data expressed as number (%). Percentages may not add up to 100 because some minor antibiotic classes were not included in this table.
Given the significant difference in the total number of smokers and nonsmokers between groups, a subanalysis was conducted on smoking and never-smoking populations. Overall, the mean number ARI encounters were not significantly different: 869/1000 p-y for all smokers and 1130 for never-smokers (P = .15). Although there was a noted difference between the total number of encounters for smokers (2532) versus never-smokers (4708), this significance was driven by the differences in encounter frequency in the control population (1445 vs 3382; P < .001). There was no significant difference in encounter frequency between smokers and never-smokers in PWH. There was also no difference between antibiotic appropriateness between smokers and never-smokers among PWH or controls (data not shown).
DISCUSSION
In this retrospective study, we aimed to assess the incidence of ARIs and antibiotic prescribing for the PWH population compared with HIV-negative controls. Because the concern of complications from infection is the most cited reason why physicians inappropriately prescribe antibiotics [11], we hypothesized that such apprehension might drive a greater degree of antibiotic prescriptions in PWH. This, however, was not borne out in our data. Overall, those with well controlled HIV appeared to access healthcare for ARIs and were prescribed antibiotics as often or less as those without HIV. However, we have identified areas for improvement in antibiotic stewardship, because approximately 35%–40% of ARI encounters resulted in inappropriate antibiotics. This is slightly higher than previous studies, which suggest closer to 25%–30% are inappropriate [5, 18, 21], although there is large variability in study time frames and population characteristics. Antibiotic appropriateness data vary widely in the literature, and, on average, within other cohorts, the percentage of inappropriate prescriptions for ARIs was reported to be approximately 60%, and as high as approximately 80% [8, 9, 15, 22, 23]. Similar to other studies, we observed that the most often used antibiotics were macrolides and fluoroquinolones [8, 13, 14, 16, 18, 19]. Our study also showed parity between PWH and without HIV in terms of the most common ICD-10 codes associated with ARIs. As in other studies, the top cited codes included the ICD-10 codes for pharyngitis and sinusitis, suggesting possible similarities in clinical presentation and diagnoses between this active duty population and other civilian cohorts [5, 14, 15].
The percentage of encounters in which antibiotics were prescribed in this study (25%–35% of encounters) was statistically significant with PWH having less antibiotics on average prescribed. This may be due to a higher percentage of encounters being conducted by Infectious Disease specialists, who may have more experience with appropriate antibiotic stewardship practices. It is a limitation of this study that the data granularity to evaluate geographic location, clinical site, and type of prescribing clinician (eg, physician/physician assistant/nurse practitioner) did not exist to allow for evaluation.
In the PWH cohort, there was a higher proportion of smokers, which may have influenced prescribers in the decision to prescribe antibiotics. When taken separately, smokers as a group (both PWH and controls) had significantly more antibiotics prescribed. However, on secondary analysis of ever-smokers versus never-smokers, the differences in antibiotic prescribing and encounter rates were driven by within-group differences in the control population and not within PWH. This occurred despite controls having fewer smokers than PWH, suggesting the ever-smoker effect was not large enough to contribute to a significant difference in antibiotic prescriptions between PWH and controls in the overall analysis, because controls still appeared to have more antibiotics prescribed despite having fewer smokers. Likewise, there were no significant differences in ever-smokers versus never-smokers as related to mean ARI encounters or differences in antibiotic appropriateness, suggesting this variable is unlikely to account for the observed differences. Unfortunately, because the smoking variable is dichotomous, the specificity of this variable is degraded. It is a limitation of our dataset that it was not possible to correlate smoking to degree of underlying, chronic lung damage due minimal ICD-10 coding among the study population that captured underlying chronic lung disease. Such chronic disease coding would have had to be done in concurrence with the captured ARI ICD-10 code, and, in practice, this was not performed frequently by prescribers. With the currently available data, it was not possible to determine whether diagnoses were coded incidentally or as the primary chief complaint of the associated visit.
There were also key differences within groups in the utilization of healthcare resources and antibiotics. People with HIV with unoptimized disease status accessed healthcare more frequently than those with optimized status as well as persons without HIV infection. This may be due to scheduled, guideline-recommended routine clinical visits, and laboratory follow ups, during which ARIs may have been incidentally diagnosed and coded, but further research is needed to determine definitive drivers.
There are several limitations to be noted. First, ICD-10 data are dependent on prescriber-chosen coding, which is subject to individual variability and coding biases. We attempted to control for this by focusing analysis using only the top 4 most common ICD-10 codes used; however, such a focus does reduce granularity in understanding the associated illness noted by the prescriber, because the top ICD-10 codes generally represented very broad symptom categories that include multiple bacterial, viral, fungal, and other etiologies. Data were also grouped on an individual encounters basis and does not distinguish between episodes of illness. Therefore, the total estimated prescribing rates per population could be overestimated if a patient had multiple antibiotics prescribed over time for the same illness.
In addition, although the inclusion of prophylactic antibiotics was attempted to be excluded by ensuring all prescriptions were associated with an ARI ICD-10, this raises the possibility that refill prescriptions for routine prophylactic prescriptions may have been included. However, because the mean CD4 count of the study population was quite high (approximately 650 cells/µL), most participants would not have qualified for prophylaxis. Early preliminary analysis using a cutoff of >60 dispensed pills only removed 8 prescription encounters from the pool, suggesting likely minimal contribution.
Such a bias is most concerning among the use of sulfa antibiotics, because these do represent an unusual choice for respiratory infections. They were seen in relatively higher percentages of use among PWH and those with poor disease control, and coincidental refilling of prophylactic medications may explain the high sulfonamide usage seen in the PWH and subgroup analysis. However, prescriptions of sulfa drugs were also seen in the control population, suggesting usages of these antibiotics not be linked to prophylaxis.
CONCLUSIONS
Overall, this large, matched-cohort study showed that inappropriate antibiotic use for ARIs was high in both groups with and without HIV. Utilizing data collected over several years, from various geographic distributions and a large group of matched patients, this study has provided the largest analysis to date of the current prescription patterns among PWH. Understanding how prescribing patterns among PWH compare to the general population may allow for more targeted education to reduce antimicrobial exposure in a group that may already have higher rates of colonization with resistant organisms [26–28]. Existing studies are small and conducted in infants [29] or deal with prescribing patterns surrounding an individual's HIV diagnosis [30]. Further research should focus on clarifying the contributing factors behind these inequities, specifically in the unoptimized populations, because there were differences in the usage of antibiotic classes between the subgroups. Given the lower percentage of antibiotic use among PWH, it may benefit overall antibiotic stewardship to elucidate the drivers of this lower rate.
Acknowledgments
Author contribution. LS and JFO conceived and designed study and performed data analysis and manuscript writing; CD, XX, and TS developed and performed data analysis; ST, XC, AN, AB, and XC performed data extraction and analysis; BKA and AG developed cohort and analyses.
Disclaimer. The views expressed are those of the authors and do not necessarily reflect the official policy or position of the Uniformed Services University of the Health Sciences, Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Brooke Army Medical Center, Walter Reed National Military Medical Center, National Institutes of Health and Department of Health and Human Services, Department of the Navy, Army, Department of Defense, nor the US Government.
Financial support. This project has been funded by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH), under Inter-Agency Agreement Y1-AI-5072, and the Defense Health Program , US Department of Defense, under Award HU0001190002. NIH Intramural Funding was provided via CAN#8010906.
Contributor Information
L Sweet, Brooke Army Medical Center, Internal Medicine, JBSA-Fort Sam Houston, Fort Sam Houston, Texas, USA.
C Daniels, Department of Criminal Justice and Criminology, St. Mary's University, San Antonio, Texas, USA.
X Xu, Department of Sociology, University of Texas San Antonio, San Antonio, Texas, USA.
T Sunil, Department of Public Health, University of Tennessee, Knoxville, Tennessee, USA.
S Topal, Infectious Disease Clinical Research Program, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA; The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, Maryland, USA.
X Chu, The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, Maryland, USA; Infectious Disease Clinical Research Program, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA.
A Noiman, The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, Maryland, USA; Infectious Disease Clinical Research Program, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA.
A Barsoumian, Brooke Army Military Medical Center, Infectious Disease Service, JBSA-Fort Sam Houston, Fort Sam Houston, Texas, USA.
A Ganesan, Infectious Disease Clinical Research Program, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA; The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, Maryland, USA; Division of Infectious Diseases, Walter Reed National Military Medical Center, Bethesda, Maryland, USA.
B K Agan, Infectious Disease Clinical Research Program, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA; The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, Maryland, USA.
J F Okulicz, Infectious Disease Clinical Research Program, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA; Brooke Army Military Medical Center, Infectious Disease Service, JBSA-Fort Sam Houston, Fort Sam Houston, Texas, USA.
References
- 1. Marston HD, Dixon DM, Knisely JM, Palmore TN, Fauci AS. Antimicrobial resistance. JAMA 2016; 316:1193–204. [DOI] [PubMed] [Google Scholar]
- 2. World Health Organization . Antimicrobial stewardship programmes in health-care facilities in low- and middle-income countries: a WHO practical toolkit. JAC-Antimicrob Resist 2019; 1:dlz072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. FTF CARB (Federal Task Force on Combating Antibiotic-Resistant Bacteria) . National action plan for combating antibiotic-resistant bacteria, 2020-2025. Washington (DC): Office of the Assistant Secretary for Planning & Evaluation, U.S. Department of Health & Human Services, 2020: 47. Available at: https://aspe.hhs.gov/reports/national-action-plan-combating-antibiotic-resistant-bacteria-2020-2025. Accessed 12 November 2021.
- 4. Centers for Disease Control and Prevention . Antibiotic use in the United States, 2018: progress and opportunities. Atlanta, GA: US Department of Health and Human Services, CDC, 2019: pp 24. [Google Scholar]
- 5. Fleming-Dutra KE, Hersh AL, Shapiro DJ, et al. Prevalence of inappropriate antibiotic prescriptions among US ambulatory care visits, 2010-2011. JAMA 2016; 315:1864–73. [DOI] [PubMed] [Google Scholar]
- 6. Ray MJ, Tallman GB, Bearden DT, Elman MR, McGregor JC. Antibiotic prescribing without documented indication in ambulatory care clinics: national cross sectional study. BMJ 2019; 367:l6461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Kobayashi M, Shapiro DJ, Hersh AL, Sanchez GV, Hicks LA. Outpatient antibiotic prescribing practices for uncomplicated urinary tract infection in women in the United States, 2002-2011. Open Forum Infect Dis 2016; 3:ofw159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Shively NR, Buehrle DJ, Clancy CJ, Decker BK. Prevalence of inappropriate antibiotic prescribing in primary care clinics within a Veterans Affairs Health Care System. Antimicrob Agents Chemother 2018; 62:e00337–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gonzales R, Malone DC, Maselli JH, Sande MA. Excessive antibiotic use for acute respiratory infections in the United States. Clin Infect Dis 2001; 33:757–62. [DOI] [PubMed] [Google Scholar]
- 10. Gonzales R, Camargo CA, MacKenzie T, et al. Antibiotic treatment of acute respiratory infections in acute care settings. Acad Emerg Med 2006; 13:288–94. [DOI] [PubMed] [Google Scholar]
- 11. King LM, Fleming-Dutra KE, Hicks LA. Advances in optimizing the prescription of antibiotics in outpatient settings. BMJ 2018; 363:k3047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Habboush Y, Guzman N. Antibiotic Resistance. In: StatPearls. Treasure Island (FL): StatPearls Publishing, 2022. [PubMed] [Google Scholar]
- 13. Kabbani S, Palms D, Bartoces M, Stone N, Hicks LA. Outpatient antibiotic prescribing for older adults in the United States: 2011 to 2014. J Am Geriatr Soc 2018; 66:1998–2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Jones BE, Sauer B, Jones MM, et al. Variation in outpatient antibiotic prescribing for acute respiratory infections in the veteran population: a cross-sectional study. Ann Intern Med 2015; 163:73–80. [DOI] [PubMed] [Google Scholar]
- 15. Bohan JG, Madaras-Kelly K, Pontefract B, et al. Evaluation of uncomplicated acute respiratory tract infection management in veterans: a national utilization review. Infect Control Hosp Epidemiol 2019; 40:438–46. [DOI] [PubMed] [Google Scholar]
- 16. Guzik J, Patel G, Kothari P, Sharp M, Ostrowsky B. UHF Outpatient ASP collaborative team. Antibiotic prescribing for acute respiratory infections in New York City: a model for collaboration. Infect Control Hosp Epidemiol 2018; 39:1360–6. [DOI] [PubMed] [Google Scholar]
- 17. Shallcross L, Beckley N, Rait G, Hayward A, Petersen I. Antibiotic prescribing frequency amongst patients in primary care: a cohort study using electronic health records. J Antimicrob Chemother 2017; 72:1818–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Chua K-P, Fischer MA, Linder JA. Appropriateness of outpatient antibiotic prescribing among privately insured US patients: ICD-10-CM based cross sectional study. BMJ 2019; 364:k5092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Truitt KN, Brown T, Lee JY, Linder JA. Appropriateness of antibiotic prescribing for acute sinusitis in primary care: a cross-sectional study. Clin Infect Dis 2021; 72:311–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Rockenschaub P, Hayward A, Shallcross L. Antibiotic prescribing before and after the diagnosis of comorbidity: a cohort study using primary care electronic health records. Clin Infect Dis 2020; 71:e50–7. [DOI] [PubMed] [Google Scholar]
- 21. Keintz MR, Abbas A, Lyden E, et al. Antibiotic prescribing for acute respiratory illnesses in persons with HIV compared with persons without HIV. Open Forum Infect Dis 2022; 9:ofac218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Klein MB, Lu Y, DelBalso L, Coté S, Boivin G. Influenzavirus infection is a primary cause of febrile respiratory illness in HIV-infected adults, despite vaccination. Clin Infect Dis 2007; 45:234–40. [DOI] [PubMed] [Google Scholar]
- 23. Kroening-Roche JC, Soroudi A, Castillo EM, Vilke GM. Antibiotic and bronchodilator prescribing for acute bronchitis in the emergency department. J Emerg Med 2012; 43:221–7. [DOI] [PubMed] [Google Scholar]
- 24. Barlam TF, Cosgrove SE, Abbo LM, et al. Implementing an antibiotic stewardship program: guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. Clin Infect Dis 2016; 62:e51–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Marcus JL, Leyden WA, Alexeeff SE, et al. Comparison of overall and comorbidity-free life expectancy between insured adults with and without HIV infection, 2000-2016. JAMA Netw Open 2020; 3:e207954. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Cotton MF, Wasserman E, Smit J, Whitelaw A, Zar HJ. High incidence of antimicrobial resistant organisms including extended spectrum beta-lactamase producing Enterobacteriaceae and methicillin-resistant Staphylococcus aureus in nasopharyngeal and blood isolates of HIV-infected children from Cape Town, South Africa. BMC Infect Dis 2008; 8:40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Rossetti B, Lombardi F, Belmonti S, et al. Nasopharingeal bacterial and fungal colonization in HIV-positive versus HIV-negative adults. New Microbiol 2019; 42:37–42. [PubMed] [Google Scholar]
- 28. Crum-Cianflone NF, Burgi AA, Hale BR. Increasing rates of community-acquired methicillin-resistant Staphylococcus aureus infections among HIV-infected persons. Int J STD AIDS 2007; 18:521–6. [DOI] [PubMed] [Google Scholar]
- 29. Ewing AC, Davis NL, Kayira D, et al. Prescription of antibacterial drugs for HIV-exposed, uninfected infants, Malawi, 2004-2010. Emerg Infect Dis 2019; 25:103–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Martin-Iguacel R, Pedersen C, Llibre JM, et al. Prescription of antimicrobials in primary health care as a marker to identify people living with undiagnosed HIV infection, Denmark, 1998 to 2016. Euro Surveill 2019; 24:1900225. [DOI] [PMC free article] [PubMed] [Google Scholar]