Perioperative Opioid Consumption and Clinical Outcomes in Surgical Patients with a Pre-Existing Opioid-Based Intrathecal Drug Delivery System

Abstract

Background:

Intrathecal Drug Delivery Systems (IDDS) have been utilized for over three decades for management of chronic pain and spasticity. Patients with IDDS may present for surgical procedures unrelated to the IDDS device, though data are limited regarding perioperative outcomes.

Materials and Methods:

This is a historical matched cohort study conducted between January 1, 2007 – December 31, 2016 of patients with an opioid-based IDDS versus matched control patients undergoing surgery excluding interventional pain procedures. Patients in the IDDS group were matched with up to two patients without an IDDS. Multivariable regression analyses were utilized to assess differences in the primary outcome of cumulative perioperative opioid consumption (i.e. intraoperative and post-anesthesia care unit [PACU] opioid consumption), and opioid consumption during the first 24 hours and 72 postoperative hours. Postoperative clinical outcomes were also assessed including escalating oxygen requirements, naloxone administration, pain-sedation mismatch, and perioperative pain service consultation.

Results:

A total of 321 surgeries were included, 112 with IDDS and 209 controls, with median (IQR) age of 57 (49, 64) years. Compared to matched controls, patients with an IDDS had greater perioperative opioid consumption (median [IQR] oral morphine milligram equivalents [OME] of 110 [60–163] vs 93 [IQR 53–142]; adjusted multiplicative increase 1.27 [95% CI 1.03–1.57], P=0.029). IDDS patients also had greater opioid consumption in the first 24 and 72 postoperative hours (multiplicative increases of 2.18 [95% CI 1.35–3.53], P=0.002, and 2.39 [95% CI 1.38–4.15], P=0.002, respectively). There were no significant differences in postoperative oxygen requirements, naloxone administration, or pain-sedation mismatch. Inpatient pain medicine consultation was more frequent in IDDS patients compared to controls (51.8% vs 6.2%, P<0.001).

Conclusion:

Patients with opioid-based IDDS received more perioperative opioids and were more likely to receive postoperative pain service consultation compared to matched controls. There were no significant differences in clinical safety outcomes, suggesting tolerance for higher opioid doses. Further research is warranted to optimize perioperative outcomes in those with IDDS.

INTRODUCTION

The Intrathecal Drug Delivery System (IDDS) is an implanted, battery-powered, and programmable pump that administers analgesics, antispasmodics, and other drugs directly into the intrathecal space at a significantly lower dose than oral or parenteral routes.^{1, 2} Prospective multicenter studies^3–5 provide evidence supporting superior analgesia, improved quality of life, and possibly improved survival in patients with IDDS compared to conventional medical management for cancer-related pain and other refractory pain disorders.^1–3 Concordant with these positive associations, IDDS is also associated with toxicity reduction by reducing systemic drug administration^3–5 as well as decreased healthcare utilization and total medical costs due to reduction in emergency room visits, hospitalization, and systemic opioid prescriptions.^6–8 Intrathecal delivery is often chosen as a route for drug delivery to reduce side effect burden for patients who cannot tolerate escalation of oral opioids. IDDS has been utilized for over three decades for management of cancer pain, non-cancer pain, and spasticity, and is becoming increasingly more prevalent in the clinical practice of pain medicine.^9–14

Additionally, patients with IDDS may undergo surgical procedures not related to IDDS implantation surgery. To date, there is limited evidence regarding the optimal perioperative management of patients with IDDS. Unfamiliarity with IDDS technology may further promote uncertainty regarding perioperative management, which has potential implications for patient outcomes.¹⁵ For example, it is unclear if patients with IDDS have different perioperative opioid requirements or experience different rates of postoperative respiratory complications than their counterparts. The availability of this data would help to inform surgical, anesthesia, and acute pain service teams on anticipated outcomes in this unique patient group, with the ultimate goal of optimizing perioperative management strategies.

In this investigation, we assess differences in perioperative opioid consumption and postoperative respiratory complications in adults with IDDS undergoing non-interventional pain surgery compared to patients without IDDS. We hypothesize that patients with an IDDS will experience increased opioid consumption, higher rates of postoperative respiratory complications, and increased utilization of inpatient pain service consultation compared to their counterparts.

MATERIALS AND METHODS

Study Design

This is a historical matched cohort study conducted at a large academic medical center. Following approval by the Institutional Review Board (IRB; Mayo Clinic, Rochester, MN), we utilized existing electronic databases to identify patients ≥ 18 years old with an implanted IDDS (present at the time of surgery) delivering intrathecal opioid medication who underwent surgical procedures between January 1, 2007 – December 31, 2016 with associated postoperative hospital admission. Furthermore, patients must have received immediate postoperative care in the postoperative care unit (PACU), as our primary outcomes included intraoperative and PACU opioid consumptions. The requirement for written informed consent was waived by the Mayo Clinic IRB. Surgical procedures were defined as any surgical procedure that required general anesthesia or monitored anesthesia care. Patients were excluded if they had previously declined research authorization for review of medical records, classified as American Society of Anesthesiologists’ (ASA) Physical Status VI (brain death), if the IDDS pump was not administering an opioid (e.g. baclofen only), or if a subsequent surgical procedure occurred <1 week from the first surgical encounter, as these patients may have inherent care complexity that could influence perioperative management and postoperative complications.

Patients in the IDDS group were matched with up to two patients without an IDDS. Controls were matched by age (+/− 10 years), sex (exact), ASA physical status score (exact), anesthetic type (general anesthesia or monitored anesthesia care; exact), type of surgical procedure (exact), and length of surgery (+/− 60 min). Within surgical types (e.g. orthopedic surgery lower extremity), patients were matched to similar surgeries to ensure comparability of surgical insults, which may influence perioperative opioid utilization and clinical outcomes. Matching was performed using a computerized algorithm which identifies the “best” match for a particular case. This is done by identifying the pool of all potential matches based on the matching variables that need to be matched exactly and then selecting the control that minimizes the sum of the absolute differences across all matching variables that do not need to be matched exactly.¹⁶

Data Collection

All patients’ records were retrospectively reviewed for demographics (age, sex), body mass index (BMI), type of procedure, ASA physical status score, Charlson score as a marker of chronic disease severity, past medical and medication history, surgery type, and total length of surgery. Perioperative opioid consumption, including intraoperative, PACU, and postoperative opioid utilization through 72 hours, were extracted using the Acute Care Datamart, an electronic data warehouse that contains detailed features on medication administration, amongst other variables, for all patients at the study institution. All administered opioids were extracted including intravenous opioids (fentanyl, hydromorphone, morphine, remifentanil, sufentanil) and oral opioids (hydrocodone, meperidine, morphine, opium/belladonna, oxycodone immediate release, oxycodone sustained release, and tramadol). These were subsequently converted to oral milligram morphine equivalents (OME) utilizing the Centers for Disease Control and Prevention opioid calculation tool¹⁷ to quantify cumulative opioid consumption, consistent with previous research.^{18, 19} As medications contained in the IDDS at the time of presentation to surgery were not available through the Acute Care Datamart, manual chart review was conducted to abstract data on total duration of IDDS therapy prior to surgery and IDDS device settings (i.e. opioid type, total opioid dose delivered, non-opioid additives, infusion delivery settings).

Outcomes of Interest

The primary outcome was cumulative perioperative opioid consumption (i.e. intraoperative + PACU opioid), with secondary outcomes of opioid consumption during the first 24 hours and 72 hours after PACU discharge, assessed by total cumulative dose and the presence of any opioid within these intervals. Intrathecal opioids being administered to the IDDS cohort through their device were not included in calculations of perioperative opioid consumption. Clinical outcomes of interest included: escalating oxygen requirements, defined by a pulse oximetry reading ≤88% with a requirement for higher doses of supplemental oxygen after PACU discharge; pain-sedation mismatch, defined as a Richmond Agitation Sedation Scale (RASS) score of −3 to −5 and numerical rating scale (NRS) score greater than 5; postoperative naloxone administration within 24 hours of the procedure end; and postoperative inpatient pain service consultation. The inpatient pain service is a specialized team consisting of physicians, advanced practitioners, and nurses who provide consultative care for evaluation and treatment of advanced acute and chronic pain needs in hospitalized patients. This service is available 24 hours per day, 365 days per year at the study institution.

Data and Statistical Analysis

Outcomes of interest were summarized by median and interquartile range for continuous variables and frequency (%) for categorical variables. Linear and logistic regression analyses for continuous and categorical outcomes, respectively, were utilized to compare outcomes for IDDS patients and matched controls. Regression analyses for opioid consumption outcomes were additionally adjusted for variables that may influence postoperative opioid consumption or clinical outcomes, including age, sex, BMI, Charlson comorbidity index, type of surgery, type of anesthetic, duration of surgery, and history of mood disorder (depression or anxiety). Since multiple surgeries are included for some IDDS patients, analyses were performed using Generalized Estimating Equations (GEE) with robust variance estimates for all outcomes with the exception of naloxone administration and pain-sedation mismatch, which were compared between groups using Fisher’s exact test. Due to skewed distributions, opioid consumption was analyzed using a log transformation (log(OME + 1)). Results are summarized by reporting the effect estimate and 95% confidence interval (CI) for IDDS versus control. For opioid consumption the effect estimate corresponds to the multiplicative increase in the geometric mean and for binary outcomes the effect estimate corresponds to the odds ratio (OR). As a post-hoc sensitivity analysis, for the primary outcome of perioperative opioid consumption, the method described by VanderWheele et al²⁰ was used to calculate the E-value quantifying the magnitude of an unmeasured confounding effect that would be required to reduce the observed multiplicative effect estimate to 1.0 (i.e. no difference between groups).

Recognizing that baseline non-IDDS-based opioid consumption may influence perioperative opioid consumption and outcomes in control patients, a pre-defined sensitivity analysis was conducted excluding 33 control patients with preoperative home opioid use. This allowed for comparison between IDDS patients with baseline intrathecal opioid consumption versus matched control patients who were opioid-naïve. Five additional IDDS patients were excluded from this analysis because they no longer had any matched controls after the application of this exclusion. As a post-hoc analysis, we assessed the indications for inpatient pain service consultation for those with an IDDS. Further, we assessed the incidence of postoperative device malfunction in those evaluated by the inpatient pain service and alterations in IDDS setting during postoperative consultation.

For all analyses, a p<0.05 was considered statistically significant. Given the issue of multiple comparisons and potential inflation of type I error rate, we employed the Benjamini-Hochberg false discovery control procedure with a false discovery rate of 5% for all adjusted outcome models. As there were no pre-existing data regarding operative interventions for patients with IDDS, a power analysis for sample size estimation for this retrospective study was deferred. However, a ten-year review was thought to provide a sizeable and clinically relevant number to draw inference regarding perioperative management and patient outcomes. All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).

RESULTS

Demographics and Baseline Characteristics

A total of 112 surgeries for 82 IDDS patients were matched to 209 surgeries in 208 control patients (Figure 1). Demographic variables and baseline clinical characteristics are presented in Table 1. Median (IQR) age in the IDDS and control cohorts was 58 (49, 66) and 57 (50, 64), respectively (P=0.832). Females comprised 68.7% in the IDDS cohort, and 69.4% in the control group (p=0.908). Patients in the IDDS cohort had modestly higher Charlson comorbidity index score of 5 (3, 9) compared to 4 (2, 6) in the control cohort (P=0.006). In addition, IDDS patients were more likely to have a history of anxiety or depression (35.7% vs 20.1%; p=0.002). There were no statistically significant differences in other demographic and baseline clinical characteristics, including body mass index, type of surgical procedure, type of anesthetic, and length of surgery.

Figure 1. CONSORT Flow Diagram.

Details on participant enrollment are provided including inclusion, exclusion, and matched control patients.

Table 1.

Demographic, clinical history and procedural characteristics of intrathecal drug delivery system (IDDS) patients and matched controls

Characteristic	IDDS (N=112)	Control (N=209†)	p-value*
Age, years	58 (49, 66)	57 (50, 64)	0.832
Sex			0.908
Male	35 (31.3)	64 (30.6)
Female	77 (68.7)	145 (69.4)
Body Mass Index, kg/m2	28.5 (23.9, 33.2)	29.2 (25.1, 34.8)	0.166
Charlson Comorbidity Index	5 (3, 9)	4 (2, 6)	0.006
History of anxiety or depression	40 (35.7)	42 (20.1)	0.002
ASA Physical Status Score			0.983
II	38 (33.9)	73 (34.9)
III	69 (61.6)	127 (60.8)
IV	5 (4.5)	9 (4.3)
Type of surgical procedure			1.00
ENT	2 (1.8)	4 (1.9)
General	23 (20.5)	44 (21.0)
Gynecology	3 (2.7)	4 (1.9)
Neurosurgery	2 (1.8)	4 (1.9)
Orthopedic	47 (42.0)	90 (43.1)
Plastics	1 (0.9)	2 (1.0)
Radiology	3 (2.7)	4 (1.9)
Thoracic	9 (8.0)	17 (8.1)
Urology	11 (9.8)	21 (10.1)
Vascular	2 (1.8)	4 (1.9)
Other	9 (8.0)	15 (7.2)
Type of Anesthesia			0.841
General	95 (84.8)	179 (85.6)
Monitored anesthesia care	17 (15.2)	30 (14.4)
Length of Surgery, minutes	105 (46, 165)	103 (44, 72)	0.919

^*Continuous variables were compared between groups using the Wilcoxon rank sum test and categorical variables were compared between groups using the chi-square test.

^†Body mass index was missing for one control.

Detailed information on IDDS therapies are provided in Table 2. The most common indication for IDDS was chronic non-cancer pain (67%). Prior to surgery, IDDS patients had received intrathecal opioids for a median of 25.5 months (IQR 4.0 – 104.5 months). Fifty-six IDDS patients (68.3%), comprising 83 surgical encounters (74.1%), had a simple continuous infusion IDDS setting, whereas 26 IDDS patients (31.7%), comprising 29 surgical encounters (25.9%), received demand bolus dosing in addition to a basal continuous infusion. Fifty-three IDDS patients (64.6%) received multiple intrathecal adjunct medications such as local anesthetic, ziconatide, and clonidine, whereas 23 IDDS patients (28.0%) received exclusively intrathecal opioids without adjunct medications. Furthermore, concomitant use of both oral opioid therapy in addition to intrathecal opioid therapy prior to surgery was reported in 71 surgical encounters (63.4%).

Table 2.

Indications, duration, and medication features for intrathecal drug delivery

Characteristic	n(%) or median (IQR)
Indication for IDDS
Chronic non-cancer pain	75 (67.0)
Cancer pain	32 (28.6)
Length of IDDS prior to surgery, months	25.5 (4.0 – 104.5)
Primary IDDS opioid
Fentanyl	6 (5.4)
Dose (mcg)	167.0 (147.5 – 420.0)
Hydromorphone	72 (64.3)
Dose (mg)	2.0 (0.7 – 3.5)
Morphine	20 (17.9)
Dose (mg)	2.3 (1.2 – 6.8)
Sufentanila	2 (1.8)
IDDS infusion setting
Continuous	83 (74.1)
Continuous + bolus	29 (25.9)
Non-opioid additives
Bupivacaine only	34 (30.4)
Clonidine only	10 (8.9)
Ziconatide only	4 (3.6)
Bupivacaine + clonidine	18 (16.1)
Clonidine + ziconatide	6 (5.4)

All percentages are based on total surgical encounters in the IDDS cohort

^aOne patient comprising two surgical encounters had a primary IDDS opioid of sufentanil, although daily IDDS dose was not reported

Perioperative Opioid Consumption

Compared to controls, patients with an IDDS consumed more cumulative perioperative opioids (median [IQR] OME of 110 [60–163] vs 93 [53–142], adjusted multiplicative increase 1.27 [95% CI 1.03–1.57], P=0.029; Table 3). Additionally, IDDS patients consumed more opioids during the first 24 hours (79 [23–279] vs 72 [0–209] OME, adjusted multiplicative increase 2.18 [95% CI 1.35–3.53], P=0.002) and 72 hours after PACU discharge (203 [41–682] vs 128 [0–381] OME, adjusted multiplicative increase 2.39 [95% CI 1.38–4.15], P=0.002). Those with an IDDS had greater odds for receiving any opioid in the first 24 hours (OR 3.18 [1.41, 7.13]; P=0.005) and 72 hours (OR 2.26 [1.05, 4.87]; P=0.038). After adjustment of significance thresholds using the Benjamini-Hochberg false discovery control procedure, these associations remained significant.

Table 3.

Clinical outcomes of intrathecal drug delivery system (IDDS) patients and matched controls^*

Outcome	IDDS (N=112)	Control (N=209)	p-value	Covariate adjusted†
				Estimate (95% C.I.)	p-value
Opioid Utilization Outcomes
Perioperative opioid, ml OME	110 (60, 163)	93 (53, 142)	0.144	1.28 (1.03, 1.59)	0.026
Opioid first 24 hours post-PACU, ml OME	79 (23, 279)	72 (0, 209)	0.010	2.23 (1.36, 3.63)	0.001
Any opioid first 24 hours post-PACU	94 (83.9)	145 (69.4)	0.011	3.30 (1.45, 7.53)	0.005
Opioid first 72 hours post-PACU, ml OME	203 (41, 682)	128 (0, 381)	0.010	2.46 (1.41, 4.32)	0.002
Any opioid first 72 hours post-PACU	94 (83.9)	152 (72.7)	0.041	2.35 (1.08, 5.11)	0.031
Clinical Outcomes
Escalating O2 Requirement	87 (77.7)	152 (72.7)	0.458
Naloxone Administration	0 (0.0)	3 (1.4)	0.554‡
Pain-Sedation mismatch	1 (0.9)	8 (3.8)	0.169‡
Pain medicine consult	58 (51.8)	13 (6.2)	<0.001

^*Continuous outcomes are summarized using median (25^th, 75^th) and binary outcomes are summarized using n (%). Due to skewed distributions, opioid requirements were analyzed using a log transformation (log(OME + 1)). Continuous outcomes were analyzed using linear regression and binary outcomes were analyzed using logistic regression. Since multiple surgeries are included for some patients, analyses were performed using Generalized Estimating Equations (GEE) with robust variance estimates for all outcomes with the exception of naloxone administration and pain-sedation mismatch which were compared between groups using Fisher’s exact test.

^†Covariates were included for age, sex, BMI, Charlson comorbidity index, history of anxiety or depression, type of surgery, type of anesthetic, and duration of surgery. For opioid requirements expressed as a continuous variable the estimate presented corresponds to the multiplicative effect for IDDS relative to control. For binary outcomes the estimate presented corresponds to the odds ratio for IDDS relative to control.

^‡Fisher’s exact test

Clinical Outcomes

IDDS patients were more likely to receive postoperative inpatient pain service consultation compared to control patients (51.8% vs 6.2%, P<0.001; Table 3). There were no significant differences in other clinical outcomes of interest, including escalating post-operative oxygen requirements, need for naloxone administration, or pain-sedation mismatch.

Sensitivity Analysis

Study results were similar after pre-defined exclusion of control patients with home opioid utilization (Table 4). IDDS patients had higher perioperative opioid consumption, higher opioid consumption 24 hours and 72 hours postoperatively, and more frequent inpatient pain service consultation compared to matched control patients. Among IDDS patients with inpatient pain service consultation (n=58), 22 (38%) consultations were reflexively placed by the surgical team given the presence of an IDDS and 62% were placed for assistance with optimizing postoperative analgesia (i.e. poorly controlled postoperative pain). All devices were formally interrogated by the inpatient pain service upon consultation, and there were no incidences of postoperative device malfunction. Daily intrathecal opioid doses were increased postoperatively by the inpatient pain service after nine surgeries (15% of pain service consultations, 8% of all IDDS surgical encounters) by a median of 20.0% (IQR 13.5% – 25.0%) from pre-operative baseline daily dose. Daily intrathecal opioid doses were reduced in only one encounter (1.7% of pain service consultations, 0.9% of all IDDS surgical encounters; 30% dose reduction). Every IDDS manipulation that occurred in our cohort was directed by the inpatient pain service. A post-hoc sensitivity analysis revealed an E-value of 1.8 for the primary outcome, which did not reveal unmeasured confounding.

Table 4.

Sensitivity analysis of primary and secondary outcomes excluding controls who reported home opioid use^*

Outcome	IDDS (N=107)	Control (N=176)	p-value	Covariate adjusted†
				Estimate (95% C.I.)	p-value
Opioid Utilization Outcomes
Perioperative opioid, ml OME	114 (60, 165)	95 (45, 145)	0.069	1.34 (1.07, 1.68)	0.011
Opioid first 24 hours post-PACU, ml OME	75 (23, 283)	58 (0, 200)	0.006	2.48 (1.49, 4.10)	<0.001
Any opioid first 24 hours post-PACU	90 (84.1)	117 (66.5)	0.004	4.21 (1.70, 10.41)	0.002
Opioid first 72 hours post-PACU, ml OME	195 (38, 652)	124 (0, 407)	0.007	2.68 (1.50, 4.79)	<0.001
Any opioid first 72 hours post-PACU	90 (84.1)	124 (70.5)	0.020	2.72 (1.20, 6.19)	0.017
Clinical Outcomes
Escalating O2 Requirement	84 (78.5)	125 (71.0)	0.259
Naloxone Administration	0 (0.0)	2 (1.1)	0.528‡
Pain-Sedation mismatch	1 (0.9)	8 (4.6)	0.160‡
Pain medicine consult	55 (51.4)	10 (5.7)	<0.001

^*The analyses of the primary outcomes were repeated after excluding 33 controls who were indicated to have preoperative home opioid use. In addition, there were 5 IDDS patients who were also excluded because they no longer had any matched controls. Continuous outcomes are summarized using median (25^th, 75^th) and binary outcomes are summarized using n (%). Due to skewed distributions, opioid requirements were analyzed using a log transformation (log(OME + 1)). Continuous outcomes were analyzed using linear regression and binary outcomes were analyzed using logistic regression. Since multiple surgeries are included for some IDDS patients, analyses were performed using Generalized Estimating Equations (GEE) with robust variance estimates for all outcomes with the exception of naloxone administration and pain-sedation mismatch which were compared between groups using Fisher’s exact test.

^†Covariates were included for age, sex, BMI, Charlson comorbidity index, history of anxiety or depression, type of surgery, type of anesthetic, and duration of surgery. For opioid requirements expressed as a continuous variable the estimate presented corresponds to the multiplicative effect for IDDS relative to control. For binary outcomes the estimate presented corresponds to the odds ratio for IDDS relative to control.

^‡Fisher’s exact test

DISCUSSION

In this matched cohort study, patients with an opioid-based IDDS were administered more opioids both during and after surgery when compared to matched controls, likely highlighting higher opioid tolerance and greater analgesic needs. Despite receiving more opioids, IDDS patients did not experience increased adverse respiratory outcomes, including escalating oxygen requirements, rescue naloxone administration, or pain-sedation mismatch. These findings overall support the safety of higher perioperative opioid utilization in IDDS patients with baseline intrathecal opioid utilization, although optimal pain management strategies remain to be defined. Further, pain management strategies should ideally be individualized to meet patient-specific analgesic requirements as guided by emotional, behavioral, and physiological responses.

The observation of higher opioid consumption in IDDS patients undergoing surgery is concordant with published animal^21–23 and human observational studies^24–28 which have demonstrated increased tolerance to antinociceptive effects of analgesics and higher analgesic requirements in those being administered intrathecal opioids. As treatment duration increases, patients with an implanted IDDS typically require multiple and frequent opioid dose escalations to maintain analgesic efficacy.²⁹ Notably, the observed gap in opioid consumption between the IDDS cohort and controls would be even greater if one were to account for intrathecal opioid doses as part of cumulative perioperative opioid utilization. However, we made the decision a priori to consider IDDS-administered opioids as the patient’s baseline usage rather than including in perioperative opioid utilization. In practical terms, most surgical and anesthesia providers will not be able to modify this baseline IDDS-based opioid utilization perioperatively. Furthermore, calculating opioid doses contained in an IDDS is often complex and challenging. In addition to intrathecal opioid, many patients with IDDS also receive non-opioid adjuvants such as local anesthetic and ziconatide. Unfortunately, there are no widely accepted guidelines to define equianalgesic conversion formulas for the translation of intrathecal opioid doses to oral formulations.

A significant concern when administering opioids and other sedative medications to patients already receiving intrathecal opioids is respiratory depression, hypoxemia, and sedation, emphasizing the need to understand the patient’s baseline cardiopulmonary status in addition to baseline sedating medications.³⁰ However, the development of tolerance to respiratory depression and sedation is a well-established phenomenon in patients receiving long-term intermittent and continuous intrathecal opioid administration.²⁵ The Polyanalgesic Consensus Conference recommendations highlight that most patients receiving IDDS opioids are already opioid-tolerant, and therefore the risk for respiratory depression in these patients is low.³¹ Our data, which provide real-world evidence for IDDS patients in the perioperative period, support this sentiment. Additional studies in non-perioperative settings have highlighted that patients on long-term IDDS opioid therapy may tolerate high doses of opioids without experiencing serious side effects.^{24, 27} As an illustrative example, a case report described a patient with bilateral metastatic lumbosacral plexopathy due to cervical cancer who was administered up to 96 mg/day of intrathecal morphine (28,800 daily OME) and did not experience sedation or respiratory depression.²⁷

Interestingly, over half of IDDS patients in this study received inpatient pain service consultation, which was eight-fold higher when compared to matched controls. This may reflect unfamiliarity of surgical, medical, and anesthesia care teams with effectively managing acute pain needs in chronic pain patients receiving intrathecal analgesics.^{32, 33} Nearly 40% of consultations were placed reflexively due to the presence of an IDDS. Remaining consults were placed for assistance in optimizing acute on chronic pain. It is notable that approximately 15% of all pain service consultations resulted in an escalation of programmed IDDS opioid doses during the postoperative period. This highlights that many IDDS patients may benefit from pain service involvement if this resource is available, though future studies are clearly warranted to define optimal management of intrathecal medications after surgery. Conversely, it is possible that increased perioperative care intensity (i.e. pain service consultations) may attenuate some of the long-term financial benefits of IDDS therapy, though this will require detailed longitudinal financial analyses. Importantly, we encountered no instances of IDDS malfunction, suggesting that postoperative device interrogation and inpatient pain service consultation may not be routinely indicated, especially with appropriate preoperative planning and in the absence of suboptimal perioperative analgesia. However, it is still important for surgical teams to recognize patients with IDDS prior to surgery to reduce risk for device damage and malfunction from surgical interference.

There are several limitations of this investigation. First, despite being the largest perioperative cohort of IDDS patients in the published literature, the sample size remains limited, which hampers outcome assessment, particularly for rare clinical outcomes. Second, differences in outcomes between IDDS patients and controls does not imply causality. Despite statistical adjustment for known confounders that may influence both exposure (i.e. IDDS) and outcomes (i.e. opioid utilization after surgery), the potential for residual confounding remains, including lack of standardization in perioperative analgesic regimens. An E-value, calculated post-hoc, suggests that after accounting for measured confounders, an unmeasured confounder associated with both IDDS and perioperative opioid consumption with an effect ratio of 1.8 would be needed to explain away observed differences in opioid consumption between groups. Given that key variables associated with postoperative opioid requirements have already been accounted for in the analysis, the authors believe that an E-value of 1.8 is large and the presence of an unmeasured confounder of this magnitude is unlikely. Third, we did not assess opioid consumption differences with incorporation of IDDS-delivered opioids, as previously discussed. Patients with any IDDS-delivered opioid were included in the IDDS cohort, though variations in IDDS-based opioid dosing and duration of therapy may influence perioperative opioid tolerance and outcomes. Fourth, we did not abstract data on history of chronic pain or substance abuse or addiction, as these data points were not reliably available in all patients. Future investigations are warranted to account for the contributions of these factors when assessing perioperative outcomes. Fifth, patients in the control group had varying degrees of preoperative opioid availability, which may influence perioperative opioid utilization. Future studies are warranted to assess differences in perioperative opioid utilization and clinical outcomes between patients receiving intrathecal opioids compared to those receiving oral opioids before surgery. Finally, we did not stratify IDDS patients based on primary indication for IDDS placement (e.g. malignancy, non-cancer chronic pain, spasticity, etc.); it is possible that outcomes may differ across these patient groups, though larger studies are warranted for this purpose.

CONCLUSION

Patients with an implanted opioid-based IDDS had higher opioid requirements during and after surgery compared to patients without an IDDS, yet both cohorts experienced similarly low rates of adverse respiratory events. These findings highlight higher opioid tolerance and greater analgesic needs in those presenting for surgery with an opioid-based IDDS. Future multicenter investigations are warranted to assess optimal analgesic strategies in patients with IDDS.

KEY POINTS.

Question:

How do perioperative outcomes of opioid consumption and postoperative respiratory complications compare between patients with an opioid-based intrathecal drug delivery system (IDDS) versus control patients without an IDDS?

Findings:

Patients with an opioid-based IDDS consumed more perioperative opioids, yet experienced similar adverse events (i.e. escalation in postoperative oxygen requirements, postoperative naloxone administration, and pain-sedation mismatch) compared to matched control patients.

Meaning:

Patients with an opioid-based IDDS have greater perioperative opioid needs with no differences in safety outcomes, suggesting tolerance for higher opioid doses.

Glossary of Terms

ASA

American Society of Anesthesiologists

BMI

Body mass index

GEE

Generalized Estimating Equations

IDDS

Intrathecal Drug Delivery Systems

IQR

interquartile range

IRB

Institutional Review Board

NRS

Numerical rating scale

OME

Oral morphine milligram equivalents

PACU

post-anesthesia care unit

RASS

Richmond Agitation Sedation Scale