Multimodal Analgesia After Posterior Fossa Decompression With and Without Duraplasty for Children With Chiari Type I

The MMA protocol (Fig 1) consisted of scheduled intravenous (IV) ketorolac alternating with scheduled per os (PO) acetaminophen and IV diazepam, transitioning to oral medications when tolerated. For refractory spasms and pain, the diazepam dose was increased. As-needed opioids were administered for severe pain (a numerical pain rating scale score of 8–10) by using 0.05 mg/kg per dose of oxycodone or 0.05 mg/kg per dose of IV morphine, PO status). Home muscle relaxants were continued. Patient-controlled analgesia was initiated at the discretion of the pain team. No epidural, preoperative, or neuropathic analgesics were used. A neurosurgeon and a pediatric hospitalist comanaging patients with Chiari collaborated to create this protocol.

This study was approved by Columbia University’s institutional review board (protocol #AAAR6867). Variables were extracted from the electronic medical record and claims processing systems for patients undergoing PFD or PFDD for CM-I from 2012 to 2017 performed by 2 surgeons using a standardized technique¹  at a single children’s hospital. Patients were excluded for being >21 years old or having concurrent procedures.

Multimodal analgesia protocol usage (MMA+) was defined as a medication record including at least 1 administered dose each of ketorolac, diazepam, and acetaminophen on POD 0 or POD 1. Patients were retrospectively stratified into 4 subgroups: PFD with no multimodal analgesia protocol usage (MMA−), PFD with MMA+, PFDD with MMA−, and PFDD with MMA+ (Fig 1). Additional independent variables included surgeon, age, sex, race, and year. The primary outcome was the postoperative opioid requirement over PODs 0 through 2 (all patients were hospitalized until POD 2, and most were discharged on POD 3). Hydromorphone, oxycodone, and morphine were converted to total oral morphine milligram Eq (MME) per kilogram. Other dependent variables included an opioid-free postoperative course (eg, patients who required between 0 and 0.1 MME/kg of opioids from PODs 0–2), discharge opioid prescriptions, antiemetic administrations, LOS, and TDC. TDC, obtained from a hospital administrative financial database, included both variable and fractional fixed costs to the institution and was inflation adjusted to 2017-equivalent US dollars.

Baseline characteristics and outcomes were compared between subgroups by using Student’s t, χ², and Wilcoxon rank tests. Two-tailed hypotheses were used with a statistical significance level of P < .05. For TDC analysis, an a priori subgroup of LOS = 2 was used because LOS is a significant driver of cost. Multiple linear regression models were used to identify predictors of opioid requirement. Data analysis was generated with SAS version 9.4 (SAS Institute, Inc, Cary, NC).

The final analytic sample included 49 cases of PFD (63%) and 29 cases of PFDD (37%). By subgroup, 16 patients were in the PFD and MMA− group, 33 were in the PFD and MMA+ group, 15 were in the PFDD and MMA− group, and 13 were in the PFDD and MMA+ group (Fig 1, Supplemental Tables 4 and 5).

Most patients required opioids (82%). However, 29% of patients in the PFDD and MMA+ group had an opioid-free course compared with 0 patients in the PFDD and MMA− group (P = .011). Of patients in the PFD and MMA+ group, 20% were opioid free.

Patients with PFD required less opioids compared with patients with PFDD. Among patients with PFD, there were no statistically significant differences between MMA+ and MMA− groups. However, among patients with PFDD, the median opioid requirement of patients in the MMA+ group was 40% that of patients in the MMA− group: 1.0 MME/kg (interquartile range [IQR] 0.5–1.8) versus 0.4 MME/kg (IQR 0.1–1.1) with P = .006. Opioid requirements were similar on POD 0, but over PODs 1 and 2, daily opioid requirements of patients in the PFDD/MMA+ group decreased compared to those of patients in the PFDD/MMA− group (Tables 1 and 2).

A multiple linear regression model was constructed to assess for independent differences in opioid requirement between patients in the MMA+ group and those in the MMA− group. After adjustment for identified confounders of surgeon and procedure type, MMA+ status was independently associated with a mean decrease in opioid requirement of 0.146 MME/kg (P = .0497) (Table 3).

No statistically significant differences were seen in secondary outcomes (LOS, TDC, and postoperative nausea and vomiting) (Supplemental Table 6).

This is 1 of the few reports regarding MMA experiences after Chiari surgery^2,3  and the first to include both PFD and PFDD. In patients receiving a regimen of scheduled acetaminophen alternating with NSAIDs and diazepam, MMA usage was independently associated with opioid use reductions in multiple regression analysis. Significant reductions in opioid requirements were also seen in univariate comparisons with MMA usage after PFDD. These results suggest opioid-sparing benefits of MMA after Chiari surgery.

In patients with PFDD, previous MMA studies revealed opioid use reductions, with 1 regimen of scheduled, alternating PO acetaminophen and ibuprofen and another using continuous ketamine infusion.^2,3  On this protocol, patients in the PFDD and MMA+ group showed significantly lower opioid requirements, and 29% had an opioid-free postoperative course.

Among patients with PFD, lower opioid requirements were seen overall. Type of surgery appeared to be the most important factor contributing to opioid use. This is not surprising because dural opening in PFDD is associated with posterior fossa irritation. To our knowledge, however, this is the first study to report this difference between PFD and PFDD surgeries.

We had hypothesized that with lower inpatient opioid requirements, prescribers would be less likely to send patients home with opioids, which could decrease potential diversions, misuse, and accidents.⁷  However, the proportion of patients receiving opioid prescriptions at discharge was not statistically different between groups despite differences in PODs 0 through 2 opioid requirements. Recent literature reveals a widespread problem of random postsurgical opioid prescription patterns, and researchers advocate for more evidence-based practices.⁷ 

Scheduled dosing,^3,4,6,8,10  IV administration,^{2,4,6,10,12,13}  and alternating acetaminophen and NSAIDs^{3–6,8–10,12}  are important reported elements of MMA protocols. Although once controversial, short-term postoperative NSAID use has shown no association with increased hemorrhage risk after craniotomy or transsphenoidal surgery in prospective randomized trials.^10,13–15  Combining acetaminophen and NSAIDs in an alternating schedule can leverage synergistic analgesic effects by double targeting elements of the cyclooxygenase pathway.⁵ 

In 1 all-oral regimen studied in a pediatric PFDD cohort, administration of acetaminophen and ibuprofen alternating every 2 hours was associated with significant reductions in opioid use, pain scores, antiemetic use, and LOS.³  We have found it difficult to consistently administer oral regimens on POD 0 in children, however, finding that IV medications circumvent common early postoperative barriers of nausea, vomiting, and sleep disruption. In a single-blind randomized controlled trial comparing PO and IV regimens of acetaminophen and NSAIDs after craniosynostosis surgery, the all-PO group revealed 3.6 times higher odds of vomiting and 14 times higher odds of nausea compared with the all-IV group.⁶  Another prospective study of scheduled, alternating IV acetaminophen and ketorolac after spinal fusion for adolescent idiopathic scoliosis revealed lower opioid use, earlier meal tolerance, and less constipation compared to opioid-based regimens.⁸ 

Hospitalists support high-value care in which comanagement teams practice opioid stewardship and use evidence-based protocols. Standardization of care requires significant support from health care administration, invoking discussions of cost. In this study, we analyzed TDC, which does not separate cost components but is easily interpretable. Although no statistically significant differences were seen in TDC or LOS, no additional costs were demonstrated.

There are limitations to this retrospective cohort analysis of a single institution’s experience with MMA after Chiari surgery. The protocol had 59% adherence, which varied over time (Supplemental Table 6). Although causes of nonadherence were not monitored, some may have included greater familiarity with opioid-focused analgesia among members of the multidisciplinary team and electronic medical record barriers (eg, no created order sets). Implementation could improve with “hard-wiring” adherence behaviors by using formalized order sets and admission to a closed postoperative unit with protocol-trained clinicians and by developing guidelines for discharge prescriptions. Study power for secondary outcomes may have been limited by the available sample in a single-institution 5-year design. Without prospective randomization, the effect and assignment of the protocol are susceptible to confounding, so multiple regression adjustment was employed. The distribution of surgeons across subgroups was not even; however, both surgeons performed the same standardized surgery, and multiple regression adjusted for confounding by surgeon. Future studies should include pain score analysis; we lacked the resources to perform one.

In this brief report, we are the first to share an experience using MMA after both PFD and PFDD procedures for CM-I. A protocol of scheduled NSAIDs alternating with scheduled acetaminophen and diazepam was associated with opioid use reductions. Further multicenter studies may better elucidate MMA’s effects on pain, cost, and enhanced recovery after surgery.