Administration of Oxycodone Hydrochloride by Target Controlled Infusion – Does the Timing Matter?

Study design and patient population

We designed a retrospective observational study adhering to the “Strengthening the Reporting of Observational Studies in Epidemiology” (STROBE) guidelines. Our study of medical records consisted of anaesthetic journals and postoperative journals from one hundred patients undergoing laparoscopic cholecystectomy.

Inclusion/exclusion criteria

Data collection consists of all laparoscopic cholecystectomies conducted at Stavanger University Hospital between September 2015 and December 2016, during which period there were no changes in operating method or anaesthetic guideline for laparoscopic cholecystectomy. Inclusion and exclusion criteria are listed in Box 1. Briefly, we included adult patients ASA 1-2 (American Society of Anaesthesiologists Physical Status Classification System) (11) who underwent elective day-case surgery using TCI and following the anaesthetic guideline for laparoscopic cholecystectomy (ref. supplement 1). Premedication consisted of oral paracetamol. OH was the only opioid in the guideline and was to be administered at the end of anaesthesia in addition to Marcain for wound infiltration. Additional medications given intraoperatively to all patients were dexamethasone, parecoxib, and ondansetron. OH was the only opioid agent given for postoperative pain treatment.

Conduct of the study

The data collection was conducted by two persons who separately reviewed anaesthetic journals and associated postoperative journals from patients after laparoscopic cholecystectomy. A total of 546 patients satisfied the study’s inclusion criteria. Of these, 399 patients were excluded due to the pre-specified exclusion criteria (Box 2). The main reason for exclusion was use of inhalation anaesthesia. Furthermore, 47 patients were excluded due to missing anaesthetic or postoperative journals, unclear writing on records or incorrect registration of total intravenous anaesthesia (TIVA) agents.

Data handling

In compliance with requirements from the Data Protection Officer at Stavanger University Hospital, no data from patient records were copied, printed or written on paper. Research data was recorded and stored anonymously on an encrypted memory chip.

Variables

Information registered for each patient was age, sex, height, weight, date and time of start of TCI, date and time of end of TCI, total dose of propofol and remifentanil, time and amount for administration of OH intra- and postoperatively, other analgesics that were administered postoperatively, and pain score (Visual Analogue Scale – VAS).

Total dose of both propofol and remifentanil given to each patient was automatically registered in the anaesthetic journal. This also applies to time of anaesthesia recorded from TCI start to TCI end. Amount of OH administered postoperatively was manually registered in the postoperative journals.

Despite long clinical experience with OH, neither the minimum effective plasma concentration nor the minimum effective analgesic concentration have been established (7). Some articles describe that maximum plasma concentration of OH is reached after about twenty minutes with intravenous (iv) administration (8,12,13). Other studies show that OH iv provides rapid pain relief after about five to eight minutes (14). Pharmacists at the pharmaceutical company Mundipharma stated that OH has maximum effect immediately after iv administration of OH (personal communication). Based on the uncertainty concerning the concentration of OH, we have in this study defined administration of OH towards the end of anaesthesia as during the last twenty minutes of TCI.

The timing of OH administration intraoperatively was split into three categories. First, we separated those patients who received OH towards the end of TCI from those who did not. Furthermore, to isolate those who received OH only towards the end of TCI (anaesthetic guideline), these were placed in a separate group (Box 3).

Statistics

The collected research data was analysed using IBM SPSS Statistics version 24, unless otherwise stated. Descriptive statistics were presented as median and interquartile range (IQR) and as mean and standard deviation (SD) for continuous data, and as counts and percentages for categorical data. Box plots were used for illustration. Median values were compared non-parametrically using Kruskal-Wallis tests.

For comparison of mean values of OH postoperatively and the combined intra- and postoperative doses of OH between groups of OH administration timing, we used regression analysis. As the outcomes were skew distributed, analyses were performed with generalised linear models with log link (15) and using robust (sandwich) estimates of standard errors. Results are presented as marginal means (i.e., average adjusted predictions (16)) with 95 per cent confidence intervals (CI), with and without adjustment for patient and operational variables. The effects of OH timing categories were tested using Wald Chi-square tests. The regression analyses were performed in Stata version 16 with function poisson applying option vce (robust), and with functions margins and test. P-values less than 0.05 were considered statistically significant.

The sample size of N = 100 was chosen based on the rule of thumb of ten observations per variable in a linear regression i.e. to allow for flexibility and adjustments in the comparisons of timing groups.

Ethical considerations

This study was approved by the Head of the Anesthesia Department, the Research Department and the Data Protection Officer at Stavanger University Hospital. According to Stavanger University Hospital’s research policy, access was granted to anaesthetic journals, associated postoperative journals and patients medication records. The Regional Committees for Medical and Health Research Ethics considered the project as a quality improvement project and therefore did not require informed consent from individual patients, ref. 603276.

Intraoperative timing of OH administration, need for postoperative OH, and combined total amounts of OH

The observed distributions of total postoperative and combined intra- and postoperative doses of OH administered for patients in different categories for timing of OH administration are displayed in Figure 3. We found no statistically significant differences in median OH administered postoperatively between the groups for timing of OH intraoperatively (Kruskal-Wallis p = 0.19). The same was the case for the combined intra- and postoperative doses of OH (p = 0.12).

Observed means (SD) of postoperatively administered OH in mg for categories 1–3 were 4.7 (3.1), 6.9 (5.1), and 4.7 (3.4), respectively; with an overall p-value for the test of equal groups of 0.032, see Table 2 which also presents confidence intervals. Adjusting for age, gender, weight, operating time and medications given during operation did not in total affect the results substantially, and the differences were still statistically significant (p = 0.012). Pairwise comparisons in the fully adjusted model indicated differences between categories 1 and 2 (p = 0.017) and between categories 2 and 3 (p = 0.042), but not between categories 1 and 3 (p = 0.83).

In total during the intra- and postoperative course, observed means (SD) of OH administered were 11.0 (2.9), 12.1 (5.5), and 9.0 (3.5) mg for patients in categories 1–3, respectively, and before adjustment these differences were statistically significant (p = 0.039). The marginal means were only slightly affected by adjusting for patient and operational variables, however the group differences were no longer statistically significant (p = 0.057).

Strengths and limitations

Our study has several strengths. Observational studies that require informed consent for use of data from medical records is reported to introduce selection bias. Significant differences between participants and non-participants may threaten the validity of results (24). In this study this was avoided through exemption from obtaining consent (approved by the Data Protection Officer). Meaning, anaesthetic journals and associated postoperative journals from all patients that met the criteria for inclusion and exclusion were included in the study. The reliability of the study is enhanced by two individuals separately reviewed anaesthetic journals and associated postoperative journals and controlled the data entry, as well as the data being recorded over a time period of 16 months, which ensured that individual factors of the anaesthesia personnel did not affect the result.

We applied, according to Stavanger University Hospital’s research policy, to the Data Protection Officer at Stavanger University Hospital for permission to access all patient medication lists. Patient records related to medicine use that could affect pain levels and outcomes were excluded from the study. Grapefruit and St. John’s wort affects the plasma concentration of OH (25). Unfortunately, in this retrospective study, it was not possible to rule out whether patients had used this preoperatively, as it is rarely recorded in a journal. Anaesthetic journals and postoperative journals registered with other anaesthetic agents or medications than the current procedure (ref. supplement 1) were excluded from the study. Some of the patients in the study had received OH orally instead of or in addition to OH intravenously. These patients were included in the sample. The oral OH dose was converted to intravenous OH dose according to guidelines at Stavanger University Hospital.

To exclude relevant interactions between TIVA agents (propofol and remifentanil) and their effect on OH, interaction analyses were conducted in collaboration with a pharmacist. The analyses were conducted in three online encyclopaedias: felleskatalogen.no, interaksjoner.no and drugs.com. In the first two, we found no relevant interactions between these agents. The interaction analyses conducted on drugs.com showed a total of 13 articles, with non-relevant results for this study. Furthermore, changed pharmacokinetics in obese patients leads to a risk of overdose (26). Obesity is defined as BMI > 30 (27). In this study analyses were done with observed weight. However, we would like to point out that the same analyses have been carried out with corrected body weight, which gave very similar results.

The study also has limitations. Information bias may have occurred due to errors related to the registration of journal data. Administered doses of propofol and remifentanil were automatically registered in anaesthetic journals. Time of administration and amount of OH were manually registered and errors may have occurred both intra- and postoperatively. Unfortunately, we cannot exclude these aspects, however, this should be taken into account for the results of the study.

Of the 499 journals that were readily assessed, 399 (80%) were excluded due to pre-specified criteria. Thus, if we assume the same would be the case for the 47 journals with unclear information, we will have lost 9–10 patients that were in the target group, which would amount to about ten per cent. We have no indication that these 9–10 patients would be systematically different from the one hundred included patients, and regard our sample as representative for the group defined by the inclusion/exclusion criteria. The relatively small sample size, especially the small number of patients in category 3 (only end), makes the study susceptible to both type I and type II errors, thus the results must be interpreted with caution and should be validated and elaborated on in a larger study.