PLoS Med plos plosmed PLOS Medicine 1549-1277 1549-1676 Public Library of Science San Francisco, CA USA 10.1371/journal.pmed.1004049 PMEDICINE-D-22-00493 Research Article Medicine and health sciencesPharmacologyDrugsAnalgesicsOpioids Medicine and health sciencesPain managementAnalgesicsOpioids Medicine and health sciencesPharmacologyDrugsOpioids Medicine and health sciencesEpidemiologyMedical risk factors Medicine and health sciencesMedical conditionsInfectious diseasesFungal diseases Medicine and health sciencesMedical conditionsInfectious diseasesSoft tissue infections Medicine and health sciencesMedical conditionsSkin diseasesSkin infections Medicine and health sciencesDermatologySkin diseasesSkin infections Medicine and health sciencesMedical conditionsInfectious diseasesSkin infections Medicine and health sciencesCardiologyEndocarditis Medicine and health sciencesEpidemiologyMedical risk factorsCancer risk factors Medicine and health sciencesOncologyCancer risk factors Biology and life sciencesPopulation biologyPopulation metricsDeath rates Opioid agonist treatment and risk of death or rehospitalization following injection drug use–associated bacterial and fungal infections: A cohort study in New South Wales, Australia Opioid agonist treatment and injecting-related bacterial and fungal infections https://orcid.org/0000-0002-5570-5556 Brothers Thomas D. Conceptualization Data curation Formal analysis Investigation Methodology Project administration Visualization Writing – original draft Writing – review & editing 1 2 3 * https://orcid.org/0000-0003-3698-7196 Lewer Dan Conceptualization Data curation Formal analysis Investigation Methodology Supervision Visualization Writing – review & editing 1 2 https://orcid.org/0000-0002-5742-4598 Jones Nicola Data curation Investigation Methodology Project administration Resources Writing – review & editing 1 https://orcid.org/0000-0003-3571-0136 Colledge-Frisby Samantha Data curation Investigation Methodology Writing – review & editing 1 https://orcid.org/0000-0001-7008-8130 Farrell Michael Methodology Project administration Resources Writing – review & editing 1 https://orcid.org/0000-0001-9864-459X Hickman Matthew Methodology Resources Writing – review & editing 4 Webster Duncan Conceptualization Methodology Supervision Writing – review & editing 3 5 https://orcid.org/0000-0002-3549-6232 Hayward Andrew Conceptualization Methodology Supervision Writing – review & editing 2 https://orcid.org/0000-0002-8513-2218 Degenhardt Louisa Conceptualization Data curation Methodology Resources Supervision Writing – review & editing 1 National Drug and Alcohol Research Centre (NDARC), UNSW Sydney, Sydney, Australia UCL Collaborative Centre for Inclusion Health, Institute of Epidemiology and Health Care, University College London, London, United Kingdom Department of Medicine, Dalhousie University, Halifax, Canada Population Health Sciences, University of Bristol, Bristol, United Kingdom Division of Infectious Diseases, Saint John Regional Hospital, Saint John, Canada Tsai Alexander C. Academic Editor Massachusetts General Hospital, UNITED STATES

I have read the journal’s policy and the authors of this manuscript have the following competing interests: In the past 3 years, LD and MF have received untied educational grant funding from Indivior and Seqirus. LD is a member of the Editorial Board of PLOS Medicine.

 * E-mail: thomas.brothers.20@ucl.ac.uk 19 7 2022 7 2022 19 7 e1004049 12 2 2022 12 6 2022 2022 Brothers et al This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Background

Injecting-related bacterial and fungal infections are associated with significant morbidity and mortality among people who inject drugs (PWID), and they are increasing in incidence. Following hospitalization with an injecting-related infection, use of opioid agonist treatment (OAT; methadone or buprenorphine) may be associated with reduced risk of death or rehospitalization with an injecting-related infection.

 Methods and findings

Data came from the Opioid Agonist Treatment Safety (OATS) study, an administrative linkage cohort including all people in New South Wales, Australia, who accessed OAT between July 1, 2001 and June 28, 2018. Included participants survived a hospitalization with injecting-related infections (i.e., skin and soft-tissue infection, sepsis/bacteremia, endocarditis, osteomyelitis, septic arthritis, or epidural/brain abscess). Outcomes were all-cause death and rehospitalization for injecting-related infections. OAT exposure was classified as time varying by days on or off treatment, following hospital discharge. We used separate Cox proportional hazards models to assess associations between each outcome and OAT exposure. The study included 8,943 participants (mean age 39 years, standard deviation [SD] 11 years; 34% women). The most common infections during participants’ index hospitalizations were skin and soft tissue (7,021; 79%), sepsis/bacteremia (1,207; 14%), and endocarditis (431; 5%). During median 6.56 years follow-up, 1,481 (17%) participants died; use of OAT was associated with lower hazard of death (adjusted hazard ratio [aHR] 0.63, 95% confidence interval [CI] 0.57 to 0.70). During median 3.41 years follow-up, 3,653 (41%) were rehospitalized for injecting-related infections; use of OAT was associated with lower hazard of these rehospitalizations (aHR 0.89, 95% CI 0.84 to 0.96). Study limitations include the use of routinely collected administrative data, which lacks information on other risk factors for injecting-related infections including injecting practices, injection stimulant use, housing status, and access to harm reduction services (e.g., needle exchange and supervised injecting sites); we also lacked information on OAT medication dosages.

 Conclusions

Following hospitalizations with injection drug use–associated bacterial and fungal infections, use of OAT is associated with lower risks of death and recurrent injecting-related infections among people with opioid use disorder.

Thomas Brothers and co-workers study health outcomes in people with injection drug use-associated infections receiving opioid agonist treatment.

 Author summary Why was this study done?

Injecting-related bacterial and fungal infections are an increasingly common cause of pain, disability, and death among people who inject drugs (PWID).

Treatment of injecting-related infections has tended to focus on antimicrobial therapy and/or surgery, without addressing underlying substance use-related needs.

Opioid agonist treatment (OAT; including methadone and buprenorphine) has been associated with decreased risks of other injecting-related health harms (including HIV infection, hepatitis C virus infection, and overdose death) and may be associated with reduced risks of recurrence after injecting-related infections.

 What did the researchers do and find?

We identified 8,943 people with opioid use disorder who were admitted to hospital with injecting-related infections in New South Wales, Australia, between 2001 and 2018.

We found that use of OAT after hospital discharge was associated with both lower risks of death and of rehospitalization with injecting-related infections.

Death and rehospitalization with injecting-related infections were common, even among study participants using OAT.

 What do these findings mean?

OAT should be considered as part of a multicomponent treatment strategy for injecting-related infections, aiming to reduce death and reinfection.

 Dalhousie University Internal Medicine Research Foundation Fellowship https://orcid.org/0000-0002-5570-5556 Brothers Thomas D. http://dx.doi.org/10.13039/501100005785 Faculty of Medicine, Dalhousie University Killam Postgraduate Scholarship, Ross Stewart Smith Memorial Fellowship in Medical Research, Clinician Investigator Program Graduate Stipend https://orcid.org/0000-0002-5570-5556 Brothers Thomas D. http://dx.doi.org/10.13039/501100000024 Canadian Institutes of Health Research CIHR-FRN 171259 https://orcid.org/0000-0002-5570-5556 Brothers Thomas D. http://dx.doi.org/10.13039/100000026 National Institute on Drug Abuse R25DA033211 https://orcid.org/0000-0002-5570-5556 Brothers Thomas D. http://dx.doi.org/10.13039/100000026 National Institute on Drug Abuse R25DA013582 https://orcid.org/0000-0002-5570-5556 Brothers Thomas D. http://dx.doi.org/10.13039/501100000272 National Institute for Health Research DRF-2018–11-ST2-016 https://orcid.org/0000-0003-3698-7196 Lewer Dan http://dx.doi.org/10.13039/501100001773 University of New South Wales Scientia PhD Scholarship https://orcid.org/0000-0003-3571-0136 Colledge-Frisby Samantha http://dx.doi.org/10.13039/501100000925 National Health and Medical Research Council PhD Scholarship https://orcid.org/0000-0003-3571-0136 Colledge-Frisby Samantha http://dx.doi.org/10.13039/501100000925 National Health and Medical Research Council 1135991 https://orcid.org/0000-0002-8513-2218 Degenhardt Louisa http://dx.doi.org/10.13039/100000026 National Institute on Drug Abuse R01DA044740 https://orcid.org/0000-0002-8513-2218 Degenhardt Louisa Australian Government Department of Health Drug and Alcohol Program TDB was supported by the Dalhousie University Internal Medicine Research Foundation Fellowship, Killam Postgraduate Scholarship, Ross Stewart Smith Memorial Fellowship in Medical Research and Clinician Investigator Programme Graduate Stipend (all from Dalhousie University Faculty of Medicine), a Canadian Institutes of Health Research Fellowship (CIHR-FRN# 171259), and through the Research in Addiction Medicine Scholars (RAMS) Program (National Institutes of Health/National Institute on Drug Abuse; R25DA033211) and the Fellow Immersion Training (FIT) Program in Addiction Medicine (National Institutes of Health/National Institute on Drug Abuse; R25DA013582). DL was funded by a National Institute of Health Research Doctoral Research Fellowship (DRF-2018–11-ST2-016). SC holds a Scientia PhD Scholarship from UNSW, Sydney and an Australian National Health and Medical Research Council (NHMRC) PhD Scholarship. LD is supported by an Australian National Health and Medical Research Council Senior Principal Research Fellowship (grant number 1135991). The OATS Study is supported by the National Institutes of Health (R01 DA044740 to LD). The National Drug and Alcohol Research Centre is supported by funding from the Australian Government Department of Health under the Drug and Alcohol Program. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data Availability Requests for data access can be submitted to the National Drug and Alcohol Research Centre (NDARC) at UNSW Sydney (ndarc@unsw.edu.au). Approval for the linkage of health data in NSW is provided under strict conditions, to protect confidentiality. Potential collaborators will be required to gain approval for data access and specific secondary analyses from the NSW Population and Health Services Research Ethics Committee. Collaborators proposing to examine research questions relating specifically to Aboriginal peoples will also be required to seek approval from the Aboriginal Health and Medical Research Council. Data may only be analysed within Australia. Introduction

Injection drug use–associated bacterial and fungal infections (e.g., skin and soft-tissue infections, endocarditis, osteomyelitis, septic arthritis, and epidural abscess) are associated with significant morbidity and mortality among people who inject drugs (PWID) and are costly for healthcare systems [1–6]. The incidence of hospitalization for injecting-related infections is increasing in many parts of the world, including Australia [7], Canada [2,8,9], South Africa [10], the United Kingdom [11], the United States of America [12–16], and India [17].

Prevention efforts to date have focused on individual-level behavior change interventions to promote more sterile drug preparation and safer drug injecting techniques. Unfortunately, these have shown mixed results [18–20] and have had limited impact on a population level [1]. This may be in part because of social and structural factors (e.g., criminalization, discrimination, lack of access to housing, harm reduction services, and supervised injection sites) that constrain the ability of PWID to inject more safely [1,21] and that push PWID away from healthcare [22]. Improved primary and secondary prevention approaches are urgently needed [1,13,22].

One promising potential intervention to prevent injecting-related bacterial and fungal infections is opioid agonist treatment (OAT; e.g., methadone or buprenorphine). For people with opioid use disorder, OAT is associated with many benefits including reduced risks of death and of viral infections including human immunodeficiency virus (HIV) and hepatitis C virus [23,24]. OAT limits opioid withdrawal symptoms, reduces reliance on illicit drug markets, and empowers PWID to inject less frequently or in a safer way [25,26]. Engagement in OAT is also associated with regular healthcare contacts where superficial infections may be treated before they progress and become more severe or spread through the bloodstream [22,27,28].

Despite these possible benefits, in many acute care hospitals, OAT is not prioritized as part of treatment planning during and after hospitalization with injecting-related bacterial and fungal infections [22,29–31]. This is represented in low rates of OAT prescribing for these patients in multiple studies from North America [29,31,32] and in qualitative studies from the UK [22]. Suboptimal access to OAT may reflect system-level issues that separate addiction care from specialized, acute medical care for infections [1,22,29,30]. In some hospitals, clinicians have tried to overcome this by establishing specialized addiction medicine consultation services [33–36] or by infectious diseases specialists prescribing OAT directly [29,37]. While OAT is known to be beneficial for other injecting-related health outcomes, there has been relatively little research on OAT and risk for injecting-related infections. A better understanding of how OAT affects outcomes after injecting-related infections could help inform treatment planning during and following hospitalization.

Prior analyses of potential benefits of OAT after hospitalization with injecting-related infections have been limited by small sample sizes with wide confidence intervals (CIs) [38,39]. Three administrative linkage cohort studies (all from US insurance claims data) have assessed associations between use of OAT and outcomes after hospitalization with injecting-related bacterial or fungal infections [39–41]. One study identified a reduced risk of death after hospitalizations with injecting-related endocarditis, but did not assess rehospitalizations [40]. A second study identified no significant effect (with wide CIs) on risk of rehospitalization after endocarditis and did not assess mortality [39]. A third identified a reduced risk of rehospitalization for skin and soft-tissue infections at 1 year [41]. Reflecting suboptimal access, use of OAT (or of naltrexone, an opioid antagonist medication used for opioid use disorder treatment in the US) was reported as 24% within 3 months following hospital discharge in the first study [40] and as 6% within 30 days following discharge in the second and third studies [39,41]. The latter 2 studies also only included information on buprenorphine use, as they did not have access to insurance claims or prescribing records for methadone.

The Opioid Agonist Treatment Safety (OATS) study is an administrative data linkage cohort study in New South Wales, Australia, which includes OAT permit records (with methadone or buprenorphine) for every person accessing OAT for opioid use disorder treatment in New South Wales from 2001 to 2018 [42,43].

Using data from the OATS study, we aimed to evaluate whether use of OAT, after discharge from hospital with injecting-related bacterial and fungal infections, is associated with decreased risk of subsequent mortality or infection-related rehospitalization.

 Methods

We conducted a retrospective cohort study using linked data from the OATS study, which has been described in detail elsewhere [42,43]. This manuscript follows the REporting of studies Conducted using Observational Routinely collected health Data statement for PharmacoEpidemiology (RECORD-PE) guidelines [44] (see S1 RECORD-PE checklist). Ethics approval was obtained from the NSW Population & Health Services Research Ethics Committee (2018/HRE0205), the NSW Corrective Services Ethics Committee, and the Aboriginal Health and Medical Research Council Ethics Committee (1400/18). We did not publish a protocol before conducting the analyses. The main analysis was prespecified before conducting the analyses, but the supplementary and sensitivity analyses were not prespecified.

Setting and data sources

The OATS cohort includes all patients prescribed methadone or buprenorphine for OAT in New South Wales, which is Australia’s most populous state and includes over one-third of all people receiving OAT in the country. Clinicians in NSW must apply to the state government and receive an authority to prescribe OAT for each participant. The database includes dates of OAT initiation and discontinuation. In NSW, there is no charge for OAT in public clinics or prisons. OAT may be prescribed and dispensed in specialized clinics or prescribed in primary care settings with medicine dispensed in community pharmacies.

All individuals with an OAT permit were linked to statewide hospitalization records, incarceration records, and vital statistics/death records between August 2001 and August 2018 using probabilistic linkage based on names, sex, date of birth, and Indigenous status, as described in the OATS study protocol [43].

 Participants

We included OATS study participants who survived at least one emergency (unplanned) hospitalization with skin and soft-tissue infection, sepsis or bacteremia, endocarditis, osteomyelitis, septic arthritis, or central nervous system infections (brain or spine abscess), identified using ICD-10 codes (see Fig 1 for study inclusion flow diagram; see S1 Table for ICD codes). We began with codes used in prior studies [8,29,39–41] and adapted our final list based on literature review and clinical input from the investigator team.

10.1371/journal.pmed.1004049.g001 Study flow diagram.

OAT, opioid agonist treatment; OATS, Opioid Agonist Treatment Safety.

To be eligible, these hospitalizations had to end with the participant discharged alive to the community (rather than transfer to another hospital) so that participants could be eligible for OAT outside the hospital (see Fig 1). This was so that the timing of potential exposure and potential outcome were aligned, to avoid problems with “immortal time bias” when participants would be unable to experience either the exposure (OAT outside of acute care hospitals) or the outcomes (rehospitalization or death) [45]. Eligible hospitalizations also had to be emergency (unplanned) admissions. We excluded routine or planned admissions (e.g., for physical therapy or diagnostic procedures) because they are unlikely to represents episodes of acute illness attributable to injecting-related infections.

 Measures Outcomes

Primary outcomes were all-cause mortality and rehospitalization with an injecting-related bacterial or fungal infection. Observed time at risk (time = 0) begins the day of discharge from participants’ earliest eligible hospitalization for injecting-related infections (see Fig 2 for graphical summary of study design). Rehospitalizations for injecting-related infections were identified using the same criteria as index hospitalizations and therefore also had to be coded as emergency (unplanned) admissions. These could occur at any time point in follow-up, so may have included both hospitalizations for new infections and for failed treatments of initial infections. Participants were censored if they were still event-free on June 29, 2018.

10.1371/journal.pmed.1004049.g002 Study design.

OAT, opioid agonist treatment.

 Primary exposure

The primary exposure was use of OAT, defined by dates with an active OAT prescription. OAT exposure was treated as time varying, by day of receipt. This means that each participant’s follow-up time was divided into exposed (on OAT) and unexposed (off OAT) episodes (ie, medication status was not necessarily constant through follow-up) [46]. We did not stratify by type of OAT (i.e., methadone or buprenorphine) as we had no hypothesis that the protective effect would differ.

Consistent with previous studies, a new OAT episode was defined as one commencing 7 or more days after the end date of a prior treatment episode [47–50]. The same definition was used for defining the end of OAT episodes, treating the 6 days following the final day of the prescription as part of the episode. The decision to incorporate the 6 days following an OAT episode into the exposure definition was originally based on consultation with clinicians and pharmacologists [50]; it has been used in previous studies by members of our group [50,51], and similar cutoffs (e.g., 3 to 6 days) have been used by others [52,53]. This approach may introduce bias by allocating outcomes to the treatment period when they actually occurred after leaving treatment; this may overestimate rates of outcomes in-treatment (on OAT) and underestimate rates of outcomes out-of-treatment (off OAT), resulting in conservate estimates of potential benefit.

 Covariates

See S1 Fig for a directed acyclic graph (DAG) describing the hypothesized relationships between OAT status, the outcomes of interest, and potential confounders. All covariates were extracted from linked hospital administrative records, unless otherwise specified.

Participant characteristics measured at the time of index hospitalization included age in years (centered to mean and standardized to units of 1 standard deviation [SD]), sex (female or not female), Indigenous status (identification as Aboriginal/Torres Strait Islander or not Indigenous), and comorbidity (defined by the count of unique ICD-10 chapters recorded in any diagnostic position for the index admission). Participant characteristics measured prior to the index hospitalization (all treated as binary) include any prior acute care hospitalizations related to poisoning or toxicity from opioids (as indicators of addiction severity; T40.0 to T40.6), alcohol (F10.0, X45, X65, Y15, T51.0), or stimulants (T40.5 T43.6), and a history of prior incarceration (which is associated with increased risk for unsafe injection practices). Dates of incarceration were derived from linked incarceration administrative records.

Characteristics of the index hospitalization include the year of admission (grouped as 2001 to 2006, 2007 to 2011, or 2012 to 2018), length of stay in days (as an indicator of initial illness severity; centered to mean and standardized to units of 1 SD), and premature patient-initiated discharge against medical advice (AMA; treated as binary). For descriptive purposes, we also classified hospitalizations by the presence of each type of injecting-related infection.

 Analysis

All analyses were conducted using R version 3.6.3. We calculated the incidence rate (with Poisson CIs) of each outcome per person-time while exposed to OAT and per person-time while unexposed to OAT during follow-up. We then described the cumulative hazard of each outcome, by OAT exposure periods, using Kaplan–Meier curves and the Simon–Makuch extension for time-varying exposures [54]. These can be interpreted as the estimated survival for patients who did not change their OAT status during follow-up. We then used Cox proportional hazards models to estimate the association between OAT and the study outcomes to generate hazard ratios, adjusting for all covariates.

Supplementary analyses

The relationship between OAT use and the outcomes (mortality or rehospitalization with injecting-related infection) may vary over time and OAT may have a larger effect closer to the time of initial hospital discharge. As such, we performed a post hoc (not prespecified) supplementary analysis to generate period-specific hazard ratios within the first year after hospital discharge, within years 2 to 3, and within years 4 to 6. We did this as an extension of our final multivariable models in the main survival analyses, adjusting for all prespecified covariates.

 Sensitivity analyses

We conducted several post hoc sensitivity analyses to test the robustness of our main analysis. First, we tested the impact of alternative OAT exposure period definitions. In our main analysis (described above), we prespecified that the 6 days following the end of an OAT episode is counted as part of the exposure. We tested whether we found similar results when reducing this exposure period to the 2 days following the OAT episode and when extending it to 10 days following the OAT episode.

We then conducted a sensitivity analysis to address a potential source of “immortal time bias” in the mortality outcome survival analysis. Immortal time occurs when, within an observation period, there is a period of time where an outcome event cannot possibly have occurred [45,55]. Because linkage between OAT record data and hospitalization data was retrospective, some participants may have had their initial hospitalization before their initial OAT record and would have been unable to experience death during this time (in other words, the fact that they have a future OAT record means they could not have died before then). We therefore constructed a new analytic sample only among participants who experienced hospitalization for injecting-related infection after their first record of OAT. We did not feel this potential issue with immortal time bias would affect the rehospitalization outcome survival analysis because participants could have experienced a rehospitalization event at any time (in this case, the fact that they have a future OAT record does not necessarily mean they could not have been hospitalized before then).

 Results Participants

We identified 8,943 participants with at least 1 hospitalization for injecting-related bacterial or fungal infections. Characteristics of the sample are summarized in Table 1. Participants were mostly men (66.0%), and median age at study entry was 38 years. Skin and soft tissue infections were present during most hospitalizations (see Table 1), and 14% of participants experienced a premature discharge “against medical advice.” Length of stay had a right-skewed distribution, with median 4 days, 75th percentile 8 days, and 99th percentile 65 days.

10.1371/journal.pmed.1004049.t001 Descriptive characteristics of the sample.
Variable Levels Total (100%)
Sample N (%) 8,943 (100%)
Participant characteristics
Age Mean ± SD 39 ± 11
Median [IQR] 38 [31 to 46]
Sex Female 3,080 (34%)
Male 5,863 (66%)
Aboriginal or Torres Strait Islander Yes 1,321 (15%)
No 7,554 (85%)
Unknown 66 (<1%)
Comorbidities1 Median [IQR] 3 [2 to 5]
1 1,183 (13%)
2 1,620 (18%)
3 1,825 (20%)
4 1,418 (16%)
5 1,040 (12%)
6+ 1,857 (21%)
Prior opioid-related hospitalization Yes 749 (8%)
No 8,194 (92%)
Prior stimulant use-related hospitalization Yes 205 (2%)
No 8,738 (98%)
Prior alcohol use-related hospitalization Yes 929 (10%)
No 8,014 (90%)
Prior experience of incarceration Yes 3,845 (43%)
No 5,098 (57%)
Index hospitalization characteristics
Year of hospitalization 2001 to 2006 2,772 (30%)
2007 to 2011 2,412 (27%)
2012 to 2018 3,809 (43%)
Distribution of infections2 Total 8,943 (100%)
Skin and soft tissue 7,021 (79%)
Sepsis/bacteremia 1,207 (14%)
Endocarditis 431 (5%)
Osteomyelitis 375 (4%)
Septic arthritis 323 (4%)
Central nervous system 69 (1%)
OAT prescription active at time of discharge Yes 4,292 (48%)
No 4,651 (52%)
Length of stay (days) Mean ± SD 8.9 ± 42
Median [IQR] 4 [2 to 8]
Discharge against medical advice Yes 1,246 (14%)
No 7,697 (86%)

1Comorbidities defined by the number of ICD-10 chapters listed during the index hospital admission.

2Percentages sum to greater than 100% because each hospitalization may have codes for multiple infection categories.

AMA, against medical advice; IQR, interquartile range; SD, standard deviation; OAT, opioid agonist treatment.

Just under half of participants (4,292; 48%) were receiving OAT at the time of discharge from their index hospitalization for injecting-related infections. Of 4,651 (52%) participants without an active OAT prescription at the time of their index hospitalization, most did not access OAT soon after discharge. For example, 199 (4%) participants initiated OAT within 1 week of hospital discharge, 410 (9%) participants initiated OAT within 4 weeks, and 706 (15%) within 12 weeks.

 Main results All-cause mortality

Out of 8,943 participants, 1,481 (17%) died during follow-up. In total, participants were followed for 65,240 person-years (median 6.56 years of follow-up per person), including 34,146 (52%) person-years exposed to OAT and 31,094 (48%) person-years unexposed. Of all participants, 2,174 (24%) remained exposed to OAT throughout the entire follow-up period, and 1,341 (15%) remained unexposed throughout.

Of the deaths, 643 (43%) occurred during an OAT exposure period, and 838 (57%) occurred while unexposed to OAT. Mortality rates were 1.88 deaths (95% CI 1.17 to 2.03) per 100 person-years exposed to OAT and 2.69 (2.51 to 2.88) per 100 person-years unexposed to OAT.

Extended Kaplan–Meier survival curves for time to death are presented in Fig 3. Cumulative hazard for death in OAT treatment versus nontreatment periods was 0.3% versus 1.2% at 30 days, 0.8% versus 2.1% at 90 days, and 2.4% versus 4.3% at 365 days.

10.1371/journal.pmed.1004049.g003 Extended Kaplan–Meier curves for time to death and time to rehospitalization among participants in the OATS study who survived an initial hospitalization with injecting-related bacterial or fungal infection.

Both analyses involve 8,943 participants. The death analysis was based on 30,667 treatment or nontreatment periods, and the rehospitalization analysis was based on 23,278 treatment or nontreatment periods. OAT, opioid agonist treatment; OATS, Opioid Agonist Treatment Safety.

Results of survival models are presented in Table 2. In the adjusted model, OAT was associated with lower hazard of all-cause death (adjusted hazard ratio [aHR] 0.63, 95% CI 0.57 to 0.70).

10.1371/journal.pmed.1004049.t002 Results of Cox regression for survival following discharge from index hospitalization with an injecting-related bacterial or fungal infection.
Variable Levels Mortality outcome Rehospitalization outcome1
Unadjusted hazard ratio (95% CI) aHR (95% CI)2 Unadjusted hazard ratio (95% CI) aHR (95% CI)2
Primary exposure
OAT Unexposed day Ref Ref Ref Ref
Exposed day 0.72 (0.64 to 0.79) 0.63 (0.57 to 0.70) 0.95 (0.89 to 1.01) 0.89 (0.84 to 0.96)
Participant characteristics
Age Years (scaled) 2.15 (2.04 to 2.26) 2.04 (1.93 to 2.17) 1.33 (1.29 to 1.37) 1.26 (1.22 to 1.31)
Sex Male Ref Ref Ref Ref
Female 0.83 (0.74 to 0.92) 0.92 (0.82 to 1.02) 1.05 (0.99 to 1.13) 1.09 (1.02 to 1.17)
Aboriginal or Torres Strait Islander No Ref Ref Ref Ref
Yes 0.72 (0.61 to 0.85) 1.02 (0.86 to 1.20) 0.95 (0.86 to 1.04) 1.00 (0.91 to 1.10)
Unknown 0.92 (0.52 to 1.62) 0.95 (0.54 to 1.69) 0.57 (0.37 to 0.88) 0.62 (0.41 to 0.96)
Comorbidities 1 Ref Ref Ref Ref
2 1.46 (1.14 to 1.89) 1.39 (1.09 to 1.78) 1.14 (1.01 to 1.28) 1.09 (0.97 to 1.23)
3 1.88 (1.49 to 2.38) 1.74 (1.38 to 2.20) 1.15 (1.02 to 1.29) 1.10 (0.98 to 1.24)
4 2.19 (1.73 to 2.79) 1.98 (1.55 to 2.51) 1.29 (1.14 to 1.46) 1.20 (1.06 to 1.36)
5 3.18 (2.50 to 4.05) 2.58 (2.03 to 3.30) 1.54 (1.35 to 1.75) 1.34 (1.18 to 1.54)
6+ 5.09 (4.09 to 6.34) 3.49 (2.79 to 4.36) 1.83 (1.63 to 2.06) 1.49 (1.32 to 1.68)
Prior opioid-related hospitalization No Ref Ref Ref Ref
Yes 1.15 (1.02 to 1.30) 1.12 (0.98 to 1.28) 1.33 (1.18 to 1.49) 1.11 (0.98 to 1.25)
Prior stimulant use-related hospitalization No Ref Ref Ref Ref
Yes 0.83 (0.66 to 1.06) 1.05 (0.82 to 1.34) 1.20 (0.96 to 1.49) 1.07 (0.85 to 1.34)
Prior alcohol use-related hospitalization No Ref Ref Ref Ref
Yes 1.09 (0.96 to 1.24) 1.06 (0.93 to 1.21) 1.31 (1.18 to 1.46) 1.16 (1.04 to 1.30)
Prior experience of incarceration No Ref Ref Ref Ref
Yes 0.76 (0.68 to 0.84) 1.00 (0.89 to 1.12) 0.99 (0.93 to 1.06) 1.02 (0.96 to 1.10)
Index hospitalization characteristics
Era of hospitalization 2001 to 2006 Ref Ref Ref Ref
2007 to 2011 1.25 (1.11 to 1.41) 0.94 (0.83 to 1.07) 1.13 (1.04 to 1.23) 1.02 (0.94 to 1.11)
2012 to 2018 1.64 (1.44 to 1.87) 0.83 (0.72 to 0.96) 1.73 (1.60 to 1.87) 1.33 (1.22 to 1.46)
Length of stay Days (scaled) 1.04 (1.02 to 1.06) 1.02 (0.99 to 1.06) 1.03 (1.01 to 1.04) 1.01 (0.99 to 1.04)
Discharge against medical advice No Ref Ref Ref Ref
Yes 0.94 (0.81 to 1.10) 1.10 (0.94 to 1.28) 1.41 (1.30 to 1.54) 1.47 (1.34 to 1.60)

1Rehospitalization with injecting-related infection.

2Fully adjusted model includes all variables listed in the table.

aHR, adjusted hazard ratio; AMA, against medical advice; CI, confidence interval; OAT, opioid agonist treatment.

 Rehospitalization for an injecting-related infection

Out of 8,943 participants, 3,653 (41%) were rehospitalized with an injecting-related bacterial or fungal infection. The distribution of infection type for these rehospitalizations was similar to the distribution during the index hospitalization. This included 2,718 (78%) hospitalizations with skin and soft-tissue infections, 556 (15%) with sepsis, 255 (7%) with endocarditis, 254 (7%) with osteomyelitis, 144 (4%) with septic arthritis, and 53 (1%) with central nervous system infections.

Participants were followed for 44,690 person-years (median 3.41 years per participant), which included 22,987 (51%) person-years exposed to OAT and 21,703 (49%) person-years unexposed. Of all 8,943 participants, 2,693 (30%) remained exposed to OAT throughout the entire follow-up period, and 2,157 (24%) remained unexposed throughout.

Of the rehospitalizations, 1,820 (50%) occurred during an OAT exposure period, and 1,833 (50%) occurred while unexposed to OAT. Incidence rates for rehospitalization with injecting-related infection were 7.92 (95% CI 7.66 to 8.29) per 100 person-years exposed to OAT, and 8.45 (8.06 to 8.84) per 100 person-years unexposed to OAT.

Extended Kaplan–Meier survival curves for time to rehospitalization are presented in Fig 3. Cumulative hazard for rehospitalization in OAT treatment versus nontreatment periods was 3.7% versus 4.3% at 30 days, 6.0% versus 7.1% at 90 days, and 12.7% versus 14.4% at 365 days.

In the adjusted model, OAT was also associated with lower hazard of rehospitalization (aHR 0.89, 95% CI 0.84 to 0.96; Table 2).

 Other analyses Supplementary analyses

In a post hoc supplementary analysis, we explored associations between OAT and mortality or rehospitalization for injecting-related infections at different points in follow-up using period-specific hazard ratios (Table 3).

10.1371/journal.pmed.1004049.t003 Period-specific aHRs for associations between OAT and all-cause mortality or rehospitalization for injecting-related infections.
Time since hospital discharge Mortality outcome Rehospitalization outcome
Within first year 0.47 (0.40 to 0.55) 0.83 (0.77 to 0.91)
Year 2 to 3 0.66 (0.54 to 0.81) 0.87 (0.76 to 0.99)
Year 4 to 6 0.76 (0.58 to 0.98) 1.10 (0.91 to 1.33)

Hazard ratios (with 95% CIs) are for OAT exposure in fully adjusted models for all covariates.

aHR, adjusted hazard ratio; CI, confidence interval; OAT, opioid agonist treatment.

 Sensitivity analyses

We conducted post hoc sensitivity analyses exploring the impact of alternative OAT exposure timing definitions. Changing our exposure definition to incorporate the 2 days following the end of the OAT episode (reduced from 6 days in the main analysis) demonstrated similar results for the association between OAT with all-cause mortality (aHR 0.51, 95% CI 0.46 to 0.57) and with rehospitalization (aHR 0.88, 95% CI 0.83 to 0.95). Extending the exposure period to incorporate 10 days following the end of the OAT episode also demonstrated similar results for mortality (aHR 0.72, 95% CI 0.65 to 0.80) and for rehospitalization (0.89, 95% CI 0.84 to 0.95).

We then conducted a post hoc sensitivity analysis for the mortality outcome, reconstructing the analytic sample only among participants who experienced hospitalization for injecting-related infection at a date following their first record of OAT. This sample was slightly smaller (n = 7,641). Compared to the main analysis, more participants (59%) had an active OAT permit at the time of discharge from their index hospitalization, and more follow-up time was exposed to OAT (59%). In the fully adjusted model in this smaller sample, OAT was also associated with lower hazard of all-cause death (aHR 0.56, 95% CI 0.51 to 0.62).

 Discussion

Among a large cohort of people with opioid use disorder who have been hospitalized with injecting-related bacterial or fungal infections, we found that OAT was associated with lower risk of mortality and of rehospitalization with these infections. Our findings of an association between OAT and lower risk of death among people with opioid use disorder are consistent with prior evidence. The magnitude of the association between OAT and lower rehospitalization risk was more modest, but we are not aware of other interventions shown to reduce risk of reinfection in this setting. Rates of death and rehospitalization remained high for this young cohort of patients, even among those exposed to OAT. Half of the sample were not prescribed OAT at the time of discharge from their initial infection-related hospitalization, and only 15% of these participants initiated OAT in the 3 months following. This suggests that OAT should be offered as part of a multicomponent treatment strategy for injecting-related infections, aiming to reduce death and reinfection.

Our findings on the benefits of OAT engagement for patients after injecting-related infection in Australia build on mixed evidence from US insurance claims databases with lower rates of OAT exposure and smaller sample sizes. One previous study, among patients with injecting-related infective endocarditis in Massachusetts, US, showed time-varying exposure to OAT or extended-release naltrexone (an opioid antagonist) after hospitalization was associated with reduced risk of death [40]. A study of patients with injecting-related infective endocarditis in a US nationwide commercial insurance claims database examined associations between buprenorphine or naltrexone within 30 days after hospital discharge and risk of rehospitalization; effect estimates were associated with wide CIs that could include both beneficial or harmful effects [39]. The sample was smaller than ours (768 participants), and less than 6% of patients were exposed to these medications during follow-up [39]. In another study analyzing patients with injecting-related skin and soft tissue infections in the same US insurance claims database, 5.5% were exposed to buprenorphine or naltrexone in 30 days following hospital discharge, and this was associated with lower risk of rehospitalization with skin and soft tissue infections at 1 year of follow-up [41]. In a retrospective chart review study of patients admitted to a Missouri, US, hospital with injecting-related bacterial or fungal infections, those who received OAT during their hospitalization and continued it at discharge were less likely to be readmitted for injecting-related infections [56]. Our findings offer more robust supportive evidence of the beneficial effects of OAT exposure following hospitalization with multiple types of injecting-related infections, a larger sample size, and higher rates of OAT exposure with more specific effect estimates.

In the present study, we identified larger effect estimates for associations between OAT use and mortality than for associations between OAT use and rehospitalization with injecting-related infections. Our findings of a large protective effect of OAT on mortality risk reduction are in keeping with prior research, including multiple observational studies showing protective effects on all-cause mortality, opioid overdose deaths, and multiple other specific causes of death (including suicide, cancer, alcohol related, and cardiovascular related) [23,57]. Future research should investigate associations between OAT and specific causes of death after hospitalization with injecting-related infections. We hypothesized several pathways through which OAT might reduce risks of recurrence of injecting-related infections, including reducing frequency of opioid injecting, improving healthcare contacts, and reducing the impacts of criminalization and violence, but we were unable to explore specific mechanisms in this study of administrative data [1,26]. People accessing OAT may still be at risk of injecting-related infections through several pathways, including ongoing injection opioid use while on OAT, suboptimal access to safe housing and harm reduction services (e.g., needle exchange and supervised consumptions sites) and by injecting stimulants. OAT is known to reduce risks of death even among people who continue to use nonmedical or criminalized opioids [58], who may still be at risk of injecting-related infections. More research is needed to understand how to further reduce risks of injecting-related infections for people both on and off OAT.

Despite the known benefits of OAT for mortality risk reduction, less than half of participants in our study had an active prescription for OAT at the time of discharge from their index hospitalization with injecting-related bacterial or fungal infections. Published rates of OAT engagement as part of discharge planning following hospitalization with injecting-related infections vary widely, including 8% in Boston, Massachusetts, US [31] and 81% in Saint John, New Brunswick, Canada [29]. Improving access to OAT requires clinical and regulatory changes, including improved education for health professionals, increasing the number of points of access and availability on-demand, facilitating multiple medication options, and decreasing out-of-pocket patient costs [59]. Infectious disease specialists should consider integrating OAT into their care of patients with injecting-related infections [29,60]. Addiction medicine physicians can be incorporated into multidisciplinary teams to help care planning for these patients [30]. The time period immediately following discharge from acute care hospitalization is a particularly dangerous time for people with opioid use disorder [61], and so hospital-based healthcare providers should offer OAT initiation and facilitate a seamless transition to ongoing, outpatient care [27,29,33,56]. Risks of death and rehospitalization remain high among people with opioid use disorder even when engaged in OAT. Addiction treatment should be considered as part of a multicomponent secondary prevention strategy that could include consideration of environmental determinants like housing and access to other harm reduction services [1,62].

Our study has some important limitations. First, the OATS study cohort does not include all people who inject opioids in NSW; only those who have accessed OAT at least once during the study period are eligible for linkage and inclusion. However, this has previously been estimated to include >75% of people with opioid use disorder in NSW [28] and, to our knowledge, our study includes the largest sample to date of people with opioid use disorder following hospitalization with injecting-related infections. Second, as this is a study of administrative healthcare data, we have no information on additional factors that may influence risk for these infections, including individual injecting practices, housing status, and access to needle exchange or supervised injection sites [1]. We had only limited information on other social determinants, aside from prior incarceration (reflecting experiences of criminalization and possible unsafe injecting technique while incarcerated) and Aboriginal or Torres Strait Islander status (reflecting experiences of cultural strengths as well as settler colonialism and structural racism) [1]. These covariates were treated as time fixed at baseline (i.e., not time varying); further research is needed to understand whether social exposures like incarceration have time-dependent effects on injecting-related infections. Third, we did not have reliable information on the dose received each day, so did not include OAT dosing information. Fourth, oral methadone and sublingual buprenorphine were the only OAT medications used in NSW during the study period, so we were unable to estimate the effects of other treatment and harm reduction modalities including slow-release oral morphine, injectable OAT (with diamorphine or hydromorphone), or the emerging practice prescribing a “safe supply” of pharmaceutical opioids to substitute for illicitly manufactured heroin or fentanyl [63].

 Conclusions

Among people with opioid use disorder following hospitalization for injecting-related bacterial or fungal infections, use of OAT is associated with lower risk of death or rehospitalization with injecting-related infections. Our findings suggest that patients with opioid use disorder and injecting-related bacterial or fungal infections can reduce their risk of death or reinfection by engaging in OAT. Clinicians, hospitals, and health systems should facilitate access to OAT and support engagement.

 Supporting information RECORD-PE, REporting of studies Conducted using Observational Routinely collected health Data statement for PharmacoEpidemiology.

(PDF)

 ICD-10 codes to define infections of interest.

(DOCX)

 DAG describing hypothesized relationships between primary exposure, covariates, and outcomes.

Figure generated with Daggity.net software. Timing of variables generally goes from the left to right. Blue circle is outcome. Green circle is exposure. Red circles are ancestors of exposures and of outcomes. White circles are adjusted variables (in this case, through study design and selection criteria). Gray circles are unobserved variables (in this case, macroenvironmental influences on risk). DAG, directed acyclic graph.

(DOCX)

Data were provided, and linkage was conducted by the NSW Ministry of Health, Centre for Health Record Linkage, and Bureau of Crime Statistics and Research. We also acknowledge the support and expertise of the OATS Study Aboriginal Advisory Group in reviewing this manuscript. The authors acknowledge the Registries of Births, Deaths and Marriages, the Coroners and the National Coronial Information System for enabling Cause of Death Unit Record File (COD URF) data to be used for this publication.

We thank the faculty and trainees in the Research in Addiction Medicine Scholars (RAMS) program and in the Fellows Immersion Training in Addiction Medicine (FIT) program for helpful feedback on the analysis plan.

Disclaimers

The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care.

 Abbreviations aHR

adjusted hazard ratio

 AMA

against medical advice

 CI

confidence interval

 DAG

directed acyclic graph

 HIV

human immunodeficiency virus

 OAT

opioid agonist treatment

 OATS

Opioid Agonist Treatment Safety

 PWID

people who inject drugs

 RECORD-PE

REporting of studies Conducted using Observational Routinely collected health Data statement for PharmacoEpidemiology

 SD

standard deviation

 References Brothers TD, Lewer D, Bonn M, Webster D, Harris M. Social and structural determinants of injecting-related bacterial and fungal infections among people who inject drugs: protocol for a mixed studies systematic review. BMJ Open. 2021 Aug 9;11(8):e049924. doi: 10.1136/bmjopen-2021-049924 34373309 Gomes T, Kitchen SA, Tailor L, Men S, Murray R, Bayoumi AM, et al. Trends in Hospitalizations for Serious Infections Among People With Opioid Use Disorder in Ontario, Canada. J Addict Med. 2021 Oct 28; Online Ahead of Print. doi: 10.1097/ADM.0000000000000928 34711742 Serota DP, Bartholomew TS, Tookes HE. Evaluating Differences in Opioid and Stimulant Use-associated Infectious Disease Hospitalizations in Florida, 2016–2017. Clin Infect Dis. 2021 Oct 1;73(7):e1649–57. doi: 10.1093/cid/ciaa1278 32886747 Meisner JA, Anesi J, Chen X, Grande D. Changes in Infective Endocarditis Admissions in Pennsylvania During the Opioid Epidemic. Clin Infect Dis. 2020;71(7):1664–70. doi: 10.1093/cid/ciz1038 31630192 Lewer D, Freer J, King E, Larney S, Degenhardt L, Tweed EJ, et al. Frequency of healthcare utilisation by adults who use illicit drugs: a systematic review and meta-analysis. Addiction. 2020 Jun;115(6):1011–23. doi: 10.1111/add.14892 31705770 Kim JH, Fine DR, Li L, Kimmel SD, Ngo LH, Suzuki J, et al. Disparities in United States hospitalizations for serious infections in patients with and without opioid use disorder: A nationwide observational study. PLoS Med. 2020 Aug 7;17(8):e1003247. doi: 10.1371/journal.pmed.1003247 32764761 Wright A, Otome O, Harvey C, Bowe S, Athan E. The Current Epidemiology of Injecting Drug Use-Associated Infective Endocarditis in Victoria, Australia in the Midst of Increasing Crystal Methamphetamine Use. Heart Lung Circ. 2018 Apr 1;27(4):484–8. doi: 10.1016/j.hlc.2017.03.166 28533098 Mosseler K, Materniak S, Brothers TD, Webster D. Epidemiology, microbiology, and clinical outcomes among patients with intravenous drug use-associated infective endocarditis in New Brunswick. CJC Open. 2020;2(5):379–85. doi: 10.1016/j.cjco.2020.05.002 32995724 Morin KA, Prevost CR, Eibl JK, Franklyn MT, Moise AR, Marsh DC. A retrospective cohort study evaluating correlates of deep tissue infections among patients enrolled in opioid agonist treatment using administrative data in Ontario, Canada. PLoS ONE. 2020 Apr 24;15(4):e0232191. doi: 10.1371/journal.pone.0232191 32330184 Meel R, Essop MR. Striking increase in the incidence of infective endocarditis associated with recreational drug abuse in urban South Africa. S Afr Med J. 2018 Jun 26;108(7):585–9. doi: 10.7196/SAMJ.2018.v108i7.13007 30004347 Lewer D, Harris M, Hope V. Opiate Injection–Associated Skin, Soft Tissue, and Vascular Infections, England, UK, 1997–2016. Emerg Infect Dis. 2019 Sep 23;23(8):1400–3. https://doi.org/10.3201/eid2308.170439 Wurcel AG, Anderson JE, Chui KKH, Skinner S, Knox TA, Snydman DR, et al. Increasing Infectious Endocarditis Admissions Among Young People Who Inject Drugs. Open Forum Infect Dis. 2016 Jul 26;3(3):ofw157. doi: 10.1093/ofid/ofw157 27800528 Barocas JA, Eftekhari Yazdi G, Savinkina A, Nolen S, Savitzky C, Samet JH, et al. Long-term Infective Endocarditis Mortality Associated With Injection Opioid Use in the United States: A Modeling Study. Clin Infect Dis. 2021 Dec 6;73(11):ciaa1346. doi: 10.1093/cid/ciaa1346 32901815 McCarthy NL, Baggs J, See I, Reddy SC, Jernigan JA, Gokhale RH, et al. Bacterial Infections Associated With Substance Use Disorders, Large Cohort of United States Hospitals, 2012–2017. Clin Infect Dis. 2020 Oct 23;71(7):e37–44. doi: 10.1093/cid/ciaa008 31907515 Cooper HLF, Brady JE, Ciccarone D, Tempalski B, Gostnell K, Friedman SR. Nationwide Increase in the Number of Hospitalizations for Illicit Injection Drug Use-Related Infective Endocarditis. Clin Infect Dis. 2007 Nov 1;45(9):1200–3. doi: 10.1086/522176 17918083 Ronan MV, Herzig SJ. Hospitalizations Related To Opioid Abuse/Dependence And Associated Serious Infections Increased Sharply, 2002–12. Health Aff. 2016 Jan 5;35(5):832–7. doi: 10.1377/hlthaff.2015.1424 27140989 Arora N, Panda PK, Cr P, Uppal L, Saroch A, Angrup A, et al. Changing spectrum of infective endocarditis in India: An 11-year experience from an academic hospital in North India. Indian Heart J. 2021 Nov 1;73(6):711–7. doi: 10.1016/j.ihj.2021.09.008 34861981 Phillips KT, Stewart C, Anderson BJ, Liebschutz JM, Herman DS, Stein MD. A randomized controlled trial of a brief behavioral intervention to reduce skin and soft tissue infections among people who inject drugs. Drug Alcohol Depend. 2021 Feb 27;108646. doi: 10.1016/j.drugalcdep.2021.108646 33677353 Stein MD, Phillips KT, Herman DS, Keosaian J, Stewart C, Anderson BJ, et al. Skin-cleaning among hospitalized people who inject drugs: a randomized controlled trial. Addiction. 2021 May;116(5):122–1130. doi: 10.1111/add.15236 32830383 Roux P, Donadille C, Magen C, Schatz E, Stranz R, Curano A, et al. Implementation and evaluation of an educational intervention for safer injection in people who inject drugs in Europe: a multi-country mixed-methods study. Int J Drug Policy. 2021 Jan 1;87:102992. doi: 10.1016/j.drugpo.2020.102992 33096364 Robertson R, Broers B, Harris M. Injecting drug use, the skin and vasculature. Addiction. 2021 Jul;116(7):add.15283. doi: 10.1111/add.15283 33051902 Harris M. Normalised pain and severe health care delay among people who inject drugs in London: Adapting cultural safety principles to promote care. Soc Sci Med. 2020 Sep 1;260:113183. doi: 10.1016/j.socscimed.2020.113183 32682207 Stone J, Degenhardt L, Grebely J, Larney S, Altice FL, Smyrnov P, et al. Modelling the intervention effect of opioid agonist treatment on multiple mortality outcomes in people who inject drugs: a three-setting analysis. Lancet Psychiatry. 2022 Apr;8(4):301–309. https://doi.org/10.1016/s2215-0366(20)30538-1 MacArthur GJ, Minozzi S, Martin N, Vickerman P, Deren S, Bruneau J, et al. Opiate substitution treatment and HIV transmission in people who inject drugs: systematic review and meta-analysis. BMJ. 2012 Oct 3;345:e5945. doi: 10.1136/bmj.e5945 23038795 Harris RE, Richardson J, Frasso R, Anderson ED. Experiences with skin and soft tissue infections among people who inject drugs in Philadelphia: A qualitative study. Drug Alcohol Depend. 2018 Jun 1;187:8–12. doi: 10.1016/j.drugalcdep.2018.01.029 29626746 Frank D. “I was not sick and I didn’t need to recover”: Methadone Maintenance Treatment (MMT) as a refuge from criminalization. Subst Use Misuse. 2018 Jan 28;53(2):311–22. doi: 10.1080/10826084.2017.1310247 28704148 Brothers TD, Fraser J, Webster D. Caring for people who inject drugs when they are admitted to hospital. CMAJ. 2021 Mar 22;193(12):E423–4. doi: 10.1503/cmaj.202124 33753366 Lewer D, Jones NR, Hickman M, Larney S, Ezard N, Nielsen S, et al. Risk of discharge against medical advice among hospital inpatients with a history of opioid agonist therapy in New South Wales, Australia: A cohort study and nested crossover-cohort analysis. Drug Alcohol Depend. 2020 Dec 1;217:108343. doi: 10.1016/j.drugalcdep.2020.108343 33122155 Brothers TD, Mosseler K, Kirkland S, Melanson P, Barrett L, Webster D. Unequal access to opioid agonist treatment and sterile injecting equipment among hospitalized patients with injection drug use-associated infective endocarditis. PLoS ONE. 2022 Jan 26;17(1):e0263156. doi: 10.1371/journal.pone.0263156 35081174 Brothers TD, Webster D. Multidisciplinary infective endocarditis care teams should address substance use disorders and harm reduction services. Open Heart. 2018 Jan 3;4(2). Available from: https://openheart.bmj.com/content/4/2/e000699.responses#multidisciplinary-infective-endocarditis-care-teams-should-address-substance-use-disorders-and-harm-reduction-services Rosenthal ES, Karchmer AW, Theisen-Toupal J, Castillo RA, Rowley CF. Suboptimal Addiction Interventions for Patients Hospitalized with Injection Drug Use-Associated Infective Endocarditis. Am J Med. 2016 May;129(5):481–5. doi: 10.1016/j.amjmed.2015.09.024 26597670 Nguemeni Tiako MJ, Hong S, Bin Mahmood SU, Mori M, Mangi A, Yun J, et al. Inconsistent addiction treatment for patients undergoing cardiac surgery for injection drug use-associated infective endocarditis. J Addict Med. 2020 Dec;14(6):e350–4. doi: 10.1097/ADM.0000000000000710 32732685 Marks LR, Munigala S, Warren DK, Liang SY, Schwarz ES, Durkin MJ. Addiction Medicine Consultations Reduce Readmission Rates for Patients With Serious Infections From Opioid Use Disorder. Clin Infect Dis. 2019 May 17;68(11):1935–7. https://doi.org/10.1093/cid/ciy924https://academic.oup.com/cid/article/68/11/1935/5142876 30357363 Braithwaite V, Ti L, Fairbairn N, Ahamad K, McLean M, Harrison S, et al. Building a hospital-based addiction medicine consult service in Vancouver, Canada: the path taken and lessons learned. Addiction. 2021 Jul;116(7):1892–900. doi: 10.1111/add.15383 33339073 Brothers TD, Fraser J, MacAdam E, Hickcox S, Genge L, O’Donnell T, et al. Implementation and evaluation of a novel, unofficial, trainee-organized hospital addiction medicine consultation service. Subst Abus. 2021;42(4):433–7. doi: 10.1080/08897077.2020.1856291 33332248 Hyshka E, Morris H, Anderson-Baron J, Nixon L, Dong K, Salvalaggio G. Patient perspectives on a harm reduction-oriented addiction medicine consultation team implemented in a large acute care hospital. Drug Alcohol Depend. 2019 Nov 1;204:107523. doi: 10.1016/j.drugalcdep.2019.06.025 31541875 Serota DP, Tookes HE, Hervera B, Gayle BM, Roeck CR, Suarez E, et al. Harm reduction for the treatment of patients with severe injection-related infections: description of the Jackson SIRI Team. Ann Med. 53 (1):1960–1968. doi: 10.1080/07853890.2021.1993326 34726095 Bassetti S, Hoffmann M, Bucher HC, Fluckiger U, Battegay M. Infections Requiring Hospitalization of Injection Drug Users Who Participated in an Injection Opiate Maintenance Program. Clin Infect Dis. 2002 Mar 1;34(5):711–3. doi: 10.1086/338876 11807681 Barocas JA, Morgan JR, Wang J, McLoone D, Wurcel A, Stein MD. Outcomes Associated With Medications for Opioid Use Disorder Among Persons Hospitalized for Infective Endocarditis. Clin Infect Dis. 2021 Feb;72(3):472–8. doi: 10.1093/cid/ciaa062 31960025 Kimmel SD, Walley AY, Li Y, Linas BP, Lodi S, Bernson D, et al. Association of Treatment With Medications for Opioid Use Disorder With Mortality After Hospitalization for Injection Drug Use–Associated Infective Endocarditis. JAMA Netw Open. 2020 Oct 14;3(10):e2016228. doi: 10.1001/jamanetworkopen.2020.16228 33052402 Barocas JA, Gai MJ, Amuchi B, Jawa R, Linas BP. Impact of medications for opioid use disorder among persons hospitalized for drug use-associated skin and soft tissue infections. Drug Alcohol Depend. 2020;215:108207. doi: 10.1016/j.drugalcdep.2020.108207 32795883 Larney S, Hickman M, Fiellin DA, Dobbins T, Nielsen S, Jones NR, et al. Using routinely collected data to understand and predict adverse outcomes in opioid agonist treatment: Protocol for the Opioid Agonist Treatment Safety (OATS) Study. BMJ Open. 2018 Aug;8(8):e025204. doi: 10.1136/bmjopen-2018-025204 30082370 Larney S, Jones N, Fiellin DA, Nielsen S, Hickman M, Dobbins T, et al. Data resource profile: The Opioid Agonist Treatment and Safety (OATS) Study, New South Wales, Australia Int J Epidemiol. 2021;49(6):1774–1775. doi: 10.1093/ije/dyaa125 33063106 Langan SM, Schmidt SA, Wing K, Ehrenstein V, Nicholls SG, Filion KB, et al. The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD-PE). BMJ. 2018 Nov 14;363:k3532. doi: 10.1136/bmj.k3532 30429167 Suissa S. Immortal Time Bias in Pharmacoepidemiology. Am J Epidemiol. 2008 Feb 15;167(4):492–9. doi: 10.1093/aje/kwm324 18056625 Therneau T, Crowson C, Atkinson E. Using Time Dependent Covariates and Time Dependent Coefficients in the Cox Model. Vienna, Austria: R Foundation; 2013 p. 31. Available from: https://cran.r-project.org/web/packages/survival/vignettes/timedep.pdf Jones NR, Shanahan M, Dobbins T, Degenhardt L, Montebello M, Gisev N, et al. Reductions in emergency department presentations associated with opioid agonist treatment vary by geographic location: A retrospective study in New South Wales, Australia Drug Alcohol Review. 2019;38(6):690–8. doi: 10.1111/dar.12976 31577058 Jones NR, Hickman M, Larney S, Nielsen S, Ali R, Murphy T, et al. Hospitalisations for non-fatal overdose among people with a history of opioid dependence in New South Wales, Australia, 2001–2018: Findings from the OATS retrospective cohort study. Drug Alcohol Depend. 2020 Oct 18;108354. doi: 10.1016/j.drugalcdep.2020.108354 33121866 Degenhardt L, Larney S, Kimber J, Gisev N, Farrell M, Dobbins T, et al. The impact of opioid substitution therapy on mortality post-release from prison: retrospective data linkage study. Addiction. 2014 Aug 1;109(8):1306–17. doi: 10.1111/add.12536 24612249 Degenhardt L, Randall D, Hall W, Law M, Butler T, Burns L. Mortality among clients of a state-wide opioid pharmacotherapy program over 20 years: Risk factors and lives saved. Drug Alcohol Depend. 2009 Nov 1;105(1):9–15. doi: 10.1016/j.drugalcdep.2009.05.021 19608355 Kimber J, Larney S, Hickman M, Randall D, Degenhardt L. Mortality risk of opioid substitution therapy with methadone versus buprenorphine: a retrospective cohort study. Lancet Psychiatry. 2015 Oct 1;2(10):901–8. doi: 10.1016/S2215-0366(15)00366-1 26384619 Cousins G, Teljeur C, Motterlini N, McCowan C, Dimitrov BD, Fahey T. Risk of drug-related mortality during periods of transition in methadone maintenance treatment: A cohort study. J Subst Abus Treat. 2011 Oct 1;41(3):252–60. doi: 10.1016/j.jsat.2011.05.001 21696913 Pearce LA, Min JE, Piske M, Zhou H, Homayra F, Slaunwhite A, et al. Opioid agonist treatment and risk of mortality during opioid overdose public health emergency: population based retrospective cohort study. BMJ. 2020 Mar 31;368:m772. doi: 10.1136/bmj.m772 32234712 Schultz LR, Peterson EL, Breslau N. Graphing survival curve estimates for time-dependent covariates. Int J Methods Psychiatr Res. 2002;11(2):68–74. doi: 10.1002/mpr.124 12459796 Yadav K, Lewis RJ. Immortal Time Bias in Observational Studies. JAMA. 2021 Feb 16;325(7):686–7. doi: 10.1001/jama.2020.9151 33591334 Marks LR, Munigala S, Warren DK, Liss DB, Liang SY, Schwarz ES, et al. A Comparison of Medication for Opioid Use Disorder Treatment Strategies for Persons Who Inject Drugs With Invasive Bacterial and Fungal Infections. J Infect Dis. 2020 Oct 1;222(Supplement_5):S513–20. doi: 10.1093/infdis/jiz516 32877547 Santo T Jr, Clark B, Hickman M, Grebely J, Campbell G, Sordo L, et al. Association of Opioid Agonist Treatment With All-Cause Mortality and Specific Causes of Death Among People With Opioid Dependence: A Systematic Review and Meta-analysis. JAMA Psychiatry. 2021 Sep 1;78(9):979–93. doi: 10.1001/jamapsychiatry.2021.0976 34076676 Stone AC, Carroll JJ, Rich JD, Green TC. One year of methadone maintenance treatment in a fentanyl endemic area: Safety, repeated exposure, retention, and remission. J Subst Abus Treat. 2020 Aug 1;115:108031. https://doi.org/10.1016/j.jsat.2020.108031 Priest KC, McCarty D, Lovejoy TI. Expanding Access to Medications for Opioid Use Disorder: Program and Policy Approaches from Outside the Veterans Health Administration. J Gen Intern Med. 2020 Dec;35(Suppl 3):886–90. doi: 10.1007/s11606-020-06266-3 33145685 Fanucchi LC, Walsh SL, Thornton AC, Nuzzo PA, Lofwall MR. Outpatient Parenteral Antimicrobial Therapy Plus Buprenorphine for Opioid Use Disorder and Severe Injection-related Infections. Clin Infect Dis. 2020 Mar 3;70(6):1226–9. doi: 10.1093/cid/ciz654 31342057 Lewer D, Eastwood B, White M, Brothers TD, McCusker M, Copeland C, et al. Fatal opioid overdoses during and shortly after hospital admissions in England: case-crossover study. PLoS Med. 2021 Oct 5;18(10):e1003759. doi: 10.1371/journal.pmed.1003759 34610017 Lennox R, Lamarche L, Martin L, O’Shea T, Belley-Côté E, Cvetkovic A, et al. The Second Heart Program—A multidisciplinary team supporting people who inject drugs with infective endocarditis: Protocol of a feasibility study. PLoS ONE. 2021 Oct 28;16(10):e0256839. doi: 10.1371/journal.pone.0256839 34710094 Bonn M, Palayew A, Bartlett S, Brothers TD, Touesnard N, Tyndall M. Addressing the Syndemic of HIV, Hepatitis C, Overdose, and COVID-19 Among People Who Use Drugs: The Potential Roles for Decriminalization and Safe Supply. J Stud Alcohol Drugs. 2020 Sep 1;81(5):556–60. https://doi.org/10.15288/jsad.2020.81.556 33028465 10.1371/journal.pmed.1004049.r001 Decision Letter 0 Turner Richard Senior Editor 2022 Richard Turner This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Submission Version 0

16 Feb 2022

Dear Dr Brothers,

Thank you for submitting your manuscript entitled "Association of opioid agonist treatment with mortality and rehospitalization following injection drug use-associated bacterial and fungal infections: linkage cohort study" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff and I am writing to let you know that we would like to send your submission for external assessment.

However, we first need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Feb 18 2022 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for full assessment.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Richard Turner, PhD

Senior Editor, PLOS Medicine

rturner@plos.org

 10.1371/journal.pmed.1004049.r002 Decision Letter 1 Turner Richard Senior Editor 2022 Richard Turner This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Submission Version 1

20 Mar 2022

Dear Dr. Brothers,

Thank you very much for submitting your manuscript "Association of opioid agonist treatment with mortality and rehospitalization following injection drug use-associated bacterial and fungal infections: linkage cohort study" (PMEDICINE-D-22-00493R1) for consideration at PLOS Medicine.

Your paper was discussed with an academic editor with relevant expertise and sent to independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to invite you to submit a revised version that addresses the reviewers' and editors' comments fully. You will appreciate that we cannot make a decision about publication until we have seen the revised manuscript and your response, and we expect to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We hope to receive your revised manuscript by Apr 08 2022 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

Please let me know if you have any questions, and we look forward to receiving your revised manuscript.

Sincerely,

Richard Turner PhD

Senior editor, PLOS Medicine

rturner@plos.org

-----------------------------------------------------------

Requests from the editors:

Please add "LD is a member of PLOS Medicine's Editorial Board", or similar, to the competing interest statement (submission form).

Please review PLOS' data policy (https://journals.plos.org/plosmedicine/s/data-availability) and adapt the data statement (submission form) to ensure that authors are not included as contacts for readers interested in inquiring about access to study data.

We ask you to adapt the title to better match journal style, and suggest: "Opioid agonist treatment and mortality and rehospitalization following injection drug use-associated bacterial and fungal infections: A linkage cohort study".

Please add a new final sentence to the "Methods and findings" subsection of your abstract, beginning "Study limitations include ..." or similar and quoting 2-3 of the study's main limitations.

After the abstract, please add a new and accessible "Author summary" section in non-identical prose. You may find it helpful to consult one or two recent research papers in PLOS Medicine to get a sense of the preferred style.

At line 96 and any other instances, please substitute "sex" for "gender" as appropriate.

Early in the Methods section (main text), please state whether or not the study had a protocol or prespecified analysis plan, and if so attach the document, referred to in the text.

Please highlight analyses that were not prespecified.

Throughout the text, please relocate reference call-outs to precede punctuation, e.g., "... similar pathways [1,25].".

Please remove the information on data sharing from the end of the main text. In the event of publication, this information will appear in the article metadata, via entries in the submission form.

In the reference list, please convert all boldface text and italics to plain text.

Please use the journal name abbreviations "PLoS ONE" and PLoS Med.".

Where appropriate, please list 6 author names rather than 3, followed by "et al.".

Please add a completed checklist for the most appropriate reporting guideline, e.g., STROBE, as an attachment, labelled "S1_STROBE_Checklist" and referred to as such in the Methods section (main text).

In the checklist, please refer to individual items by section (e.g., "Methods") and paragraph numbers, not by line or page numbers as these generally change in the event of publication.

Comments from academic editor:

1. I see that time at risk begins immediately after discharge. Are "bounce backs" (i.e., readmission within a short time interval, e.g. 24 hours, 7 days, etc) therefore considered new hospitalizations? And if so, is there any reason to suspect this may bias the estimated associations away from the null? Perhaps a sensitivity analysis might be reassuring. Either way, please include some discussion of this in the text.

2. The authors state that the six days following an OAT episode is counted as part of the exposure. How was the decision made to specify the threshold at six days? Please describe in in the text and/or state whether or not a sensitivity analysis (e.g., with alternative threshold periods) is necessary.

3. Relatedly, the authors describe how person-time exposure was counted as "while exposed to OAT" vs. "while unexposed to OAT". Is there any need to account for any lag effects? For instance, suppose a study participant is unexposed to OAT and during this time engages in injection drug use. Then they re-engage in OAT and are therefore counted as exposed. But then 2 days later, due to the injection drug use behavior from several days before, this individual becomes sicker and requires hospitalization. They would be hospitalized but would be counted as exposed.

4. I would like to see the text provide more speculation about the magnitude of the difference between the effect of exposure on mortality vs. the effect of the exposure on rehospitalization. The large difference, and the fact that the mortality benefit occurs instantaneously (i.e., after discharge), makes me worry about potentially unobserved confounding.

5. Was the effect of exposure similar for Aboriginal or Torres Strait Islanders vs. non-Aboriginal or Torres Strait Islanders?"

Comments from the reviewers:

*** Reviewer #1:

I mostly confine my remarks to statistical aspects of this paper. These were generally very well done, but I do have a few comments and suggestions.

p. 2 line 32 What is the number after the ± sign? An sd? A CI? Something else? Please specify.

p. 4 line 48-49 "Many parts of the world" ought to include some places in Africa, Asia, and less developed countries generally. If there's no data on those countries then maybe say "in many developed countries" or something like that.

p. 6-7 line 104-5 Maybe I am missing something but are the "eligible index hospitalizations" the control group? The linking mechanism is described earlier, but then this group doesn't seem to be described in the section on particiapants.

p. 8 line 139 Centering age is fine, but I am not a fan of scaling variables to 1 SD. I know some statisticians do recommend it, but I think it makes things a little less clear. For age (as here) a year is a year, but an SD in one study is different from an SD in another study. I won't insist on changing this, but I do recommend using age. (Same comment for LOS, below).

p, 9 line 148-9 Why group admission? Categorizing continuous variables is not recommended. It increases both type 1 and type 2 error and introduces a kind of "magical thinking" that something big happens at the cutpoints. Instead, leave year as a number and use splines to investigate nonlinearity. (Splines could also be used with age).

Table 1 - LOS is clearly very skew. Giving median and IQR is fine, but a density plot of this variable would give more insight. What accounts for the huge SD? Is it a single outlier (someone with 500 days or something) or is it more general right skew? A density plot or quantile plot would reveal this.

Peter Flom

*** Reviewer #2:

In this study, Brothers et all use data from the Australia OAT registry to example how the exposure of recent OAT relates to 2 outcomes: death and hospitalization. The authors are clearly very informed about the subject area, and I know their work well. My major issue with this is that the authors do not highlight as much why this specific registry is interesting and notable. There is a tension between "We know this" and "We don't know this." After reading this introduction, it is a hard sell to me that this work needs to be done. But, as you read the methods more, you realize how special this data set is because it is a closed system with really granular data. I say closed system not exactly knowing how many people leave New South Wales to go live in other places in Australia. Overall I think it is less than, say, people moving from Massachusetts to Rhode Island or New Hampshire. We would never be able to publish this type of data in the US. Additionally, 15% of people in this study are aboriginal. And I am not sure how comfortable the authors feel about diving into the details about disparities in OAT to aboriginal people, but this seems like the right time to use this data to make that statement. Yes, it will not work towards the idea of "generalizability" because not all readers will interact with aboriginal people. But it is hard to find this type of database with high percentages of native americans, and it could at least be the start to a conversation about why certain people are left out of the conversation. It looks like in the adjusted analysis the benfit of OAT goes away for both outcomes. So, I think this deserves to be published but the authors need to work on convincing me to read this paper. And the angle to sell this paper is the unique data set and the health disparities angle for sure. I recommend reframing the paper. Benefits of OAT in PWID Not Seen In Aboriginal People: Why?

Introduction:

Concise and overall well done. In the first paragraph I might say specifically what the primary and secondary prevention methods are (safe injection sites, early identification of OUD, decreased barriers to OAT access). On the revisions it makes sense to reframe talking about major issues with known data that this paper/this data set can help answer. If you look at work by leaders like Laura Marks or Sim Kimmel, I think in the have people who are "lost to follow up" or they discuss issues with the linkage of data. I believe it has been part of my work as well, because I can only tell if people die in MA. I also don't think you can say OAT has not been prioritized without saying there are some places who have prioritized OAT, but many places have not. Do you think there is limited evidence that OAT prevents against death and rehospitalization? I think that is a stretch. I think there is more and more. And maybe that is where my disconnect lies with the paper. I think there is a lot of evidence about OAT protecting against hospitalization and death. I like your comment about limited sample size. It gives you another avenue to say this is a large group of people. Also please highlight more that this cohort includes incarcerated people (or, is it history of incarceration?). That is an important point and distinction

Methods:

-What is an unplanned hospitalization? You mean if they were delivering a baby, and happened to note there was cellulitis, they did not get included? If they had a planned hospitalization and found to have some sort of infection, why would they not be included? Maybe there is a better word for "unplanned" that can help me understand? I am curious about the 776 planned hospitalizations. I am guessing they were planned AND there was no infection?

-I understand why you are selecting only outpatient OAT, but I think you need to be clearer in the intro about this. Also, if someone is going to a rehab on OAT, I would think they are sicker and more likely to be rehospitalizated. That is 1/10th of your sample (983 people). Is it possible to look at rehospitalization and death in the 983 people who did not go to community? Can you include this in your conclusions? You are selecting a healthier population over all.

-mycobaterial infections excluded? We have peopel with injection related mycobacterium.

-Was the primary exposure before hospitalization, during hospitalization, or could be after hospitalization? Did they meet the primary exposure if they got just 1 day of meds? --> (notes from later. Ok, I see in the table it is y/n prescription active at discharge; why do you need time varying based on day of receipt? not sure what that means.)

-Is it possible to go one step back from the 40,000 people and let me know how many people were in the OATs cohort all together onFigure 1

-History of incarceration may be linked to high risk injection practices, but I think current incarceration less so. I just think you need to justify/clarify this statement more.

-I might be confused because I don't know Poisson well, but if Mr Jones is on and off OAT, does the days he is on OAT qualify for one of the exposures and the days he is off qualify him for the other? Or does he just fall into one bucket.

Results

-I find the distribution of the infections fascinating. Such a high amount of hospitalizations for skin and soft tissue infections. Would be great if you could add somehting letting me know skin and soft tissue + something else.

-lines 174 further confuses me. Is it TIME of index hospitalization? Or DISCHARGE? I need more clarity of the exposure of interest. I think I am confused becayse you are using it in 2 ways. Y/N, and then how much exposure for the KM curve.

-Only 43% had prior experience of incarceration? How did you collect this data? Can you add into methods? Self report or pulled from admin data?

-What do. you think about changing your KM curves into stratified analysis of Aboriginal and note. The Aboriginal peopel are not seeing the benefit. I feel like that is the biggest statemennt you can make. There is benefit but it is not equal.

Discussion

-Rework discussion to be around equity. You found something, but it was not equitable benefit. That I believe is how this paper really makes changes and advances the discussion

-LIne 285-288. Not super relevant to your discussion. You excluded these people

-Discuss the other limitations, like excluding people who go to rehab from your analysis.

-Limiations of how places without high percentages of aboriginal people may not feel this applies, but it is about "MINORITIZED" communities overall, not necessarily WHY they are minoritized

*** Reviewer #3:

This manuscript describes the association of opioid agonist treatment (methadone and buprenorphine) with mortality and re-hospitalization following injection drug use-associated bacterial and fungal infections in New South Wales, Australia between 2001 and 2018. The authors find that current OAT use is associated with both decreased rehospitalization and mortality in adjusted models. Hospitals, public health practitioners, infectious disease and addiction clinicians all seek to improve care for people with injection-related infections and this manuscript adds to the evidence. The manuscript is clear and well written and makes an important contribution to the field. I have several suggestions however for the authors to consider which I believe would improve the strength of this manuscript.

1. It seems that Individuals in the OATS cohort who experienced a designated hospitalization for infection were eligible for inclusion regardless of whether OAT was received prior to the hospitalization or after the hospitalization. Though this may be a small number of individuals, it is possible that someone could have received OAT only several years after experiencing a designated hospitalization and still be included in the cohort. This represents a form of immortalized time bias --- individuals had to have survived the period after the hospitalization in order to be included in the cohort. This potential bias however would be toward the null. To protect against biased estimates of effect, the authors could consider only including individuals with any OAT receipt prior to the hospitalization in the cohort.

2. Figure 2 would be more accurate if it also designated OAT receipt.

2. The study includes differential follow up time after the hospitalization. The analysis includes methods to address this but assumes that the impact of OAT is the same throughout the entire follow up period. Given that the relationship to the infection and the outcome varies over time (ie. infection may be less relevant to the outcomes years out), it would be instructive to report time varying hazards for example (ie. year 1, year 2-3, and year 4-6).

3. Figure 3 would be improved with life tables. Also, the average follow up time is 6 years, but the figure only include 3 years of follow up?

4. This may be beyond the scope of this study, but given the limited evidence in the literature, it may be instructive to also report how the association between OAT and mortality and rehospitalization varies by type of infection (e.g. skin and soft tissue vs more serious infections). Such an analysis would be instructive but could also represent its own manuscript in the future.

***

Any attachments provided with reviews can be seen via the following link:

[LINK]

 10.1371/journal.pmed.1004049.r003 Author response to Decision Letter 1 Submission Version 2

9 May 2022

Submitted filename: R1_reviewerresponse_OATS_afterhospital_PLOSMed_220502.docx

 10.1371/journal.pmed.1004049.r004 Decision Letter 2 Turner Richard Senior Editor 2022 Richard Turner This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Submission Version 2

5 Jun 2022

Dear Dr. Brothers,

Thank you very much for re-submitting your manuscript "Opioid agonist treatment and risk of death or rehospitalization following injection drug use-associated bacterial and fungal infections: a linkage cohort study" (PMEDICINE-D-22-00493R2) for consideration at PLOS Medicine.

I have discussed the paper with our academic editor and it was also seen again by three reviewers. I am pleased to tell you that, once the remaining editorial and production issues are fully dealt with, we expect to be able to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We hope to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

Please let me know if you have any questions, and we look forward to receiving the revised manuscript.   

Sincerely,

Richard Turner, PhD

Senior Editor, PLOS Medicine

rturner@plos.org

------------------------------------------------------------

Requests from Editors:

We ask you to remove or reword "... will be reviewed by the OATS investigator team" from the data statement. This seems to suggest that the author group can prevent data release to interested parties, which would not be consistent with PLOS' data policy (https://journals.plos.org/plosmedicine/s/data-availability).

We suggest incorporating "in Australia" in the full title.

At line 34 and any similar instances, please make that "39 years". Incidentally, we notice that mean and median ages are both quoted at different points in the ms, and you may wish to make this consistent.

Around line 41, please adapt the summary of limitations so that this extends to only the final sentence of the "Methods and findings" subsection of the abstract.

Please avoid "reduced" and "decreased" (risk of death), e.g., at lines 48, 397-399 and 492, in favour of "lower", for example, to avoid implying causality.

At line 145-146, please substitute the label for the attached checklist.

In the reference list, please abbreviate "Drug Alcohol Depend." consistently.

Noting reference 9, please use the journal name abbreviation "PLoS ONE".

Please remove "[Internet]" from references 46, 57 and any others.

Is reference 61 missing the year of publication?

Comments from Reviewers:

*** Reviewer #1:

The authors have addressed my concerns and I now recommend publication

*** Reviewer #2:

The authors have done a great job with all of the reviewers comments. Without hesitation, I think this paper is worthy of publication!

*** Reviewer #3:

I thank the authors for their thorough and thoughtful revision.

My questions and concerns have been thoroughly addressed and I support publication of this important contribution!

***

Any attachments provided with reviews can be seen via the following link:

[LINK]

 10.1371/journal.pmed.1004049.r005 Author response to Decision Letter 2 Submission Version 3

9 Jun 2022

Submitted filename: R2_response_OATS_afterhospital_PLOSMed_220609.docx

 10.1371/journal.pmed.1004049.r006 Decision Letter 3 Turner Richard Senior Editor 2022 Richard Turner This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Submission Version 3

12 Jun 2022

Dear Dr Brothers, 

On behalf of my colleagues and the Academic Editor, Dr Tsai, I am pleased to inform you that we have agreed to publish your manuscript "Opioid agonist treatment and risk of death or rehospitalization following injection drug use-associated bacterial and fungal infections: a cohort study in New South Wales, Australia" (PMEDICINE-D-22-00493R3) in PLOS Medicine.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process. 

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper. 

Sincerely, 

Richard Turner, PhD 

Senior Editor, PLOS Medicine

rturner@plos.org