Patients
We included twenty-nine patients admitted during a nine-month period to the Palliative Care Unit at the University Hospital in Trondheim. The inclusion criteria were; verified malignant disease, expected survival time less than 6 months, scheduled morphine treatment started at least three days prior to inclusion, stable scheduled doses of morphine for a minimum of three days and age more than eighteen years. The exclusion criteria were; planned hospitalisation less than three days and lack of ability to communicate (e.g. dementia, deafness).
All patients gave their written informed consent before inclusion. The study was conducted according to the guidelines of the Helsinki declaration. The Regional Committee for Medical Research Ethics, Health Region IV, Norway, approved the study.
Study design
Inclusion
The patients were included in the study within three days after admission to the Palliative Care Unit. Each patient was followed for four days. Patients readmitted to the Palliative Care Unit were allowed to a new trial period identical to the first trial period. No patients were included in more than three trial periods.
The patients' age, gender, primary malignant diagnosis, presence of metastasis, and other medications were registered. Morphine treatment during the last 24 hours was registered with respect to route of administration, morphine formulation, scheduled dose and consumption of rescue morphine for breakthrough pain. The patients' functional status was assessed using the Karnofsky performance status score [15].
Blood samples
Blood samples were obtained each day during the trial period. The samples were obtained at the same time each day during the routine morning round for collecting blood samples.
Observations
In order to observe if the patients were studied during stable treatment conditions the scheduled morphine dose, rescue morphine consumption and route of administration were registered each study day. The use of other medications was also registered daily. Pain, nausea and sedation were assessed at day study two, three and four during the trial period using a 5 category verbal rating scale (VRS) score ranging from no to very severe. All symptoms were assessed for the last 24 h.
Analyses
The blood samples were placed in EDTA tubes until separated by centrifugation (3000 rpm, ten minutes) and stored at -85°C until analysed. All samples were analysed for serum concentrations of morphine, M6G and M3G applying liquid chromatography mass spectrometry [16]. The limits of detection were for morphine 0,35 nmol/l and for M6G and M3G 2,2 nmol/l. The analytical coefficients of variation obtained in quality control samples (CVAnalytical) were for morphine 3,0%, for M6G 5,5% and for M3G 7,0%. The analytical coefficients of variation were determined at 100 nmol/l for morphine and 1000 nmol/l for M6G and M3G. Serum values of creatinine concentrations, alanin aminotransferase activities (ALAT), aspartat aminotransferase activities (ASAT) and albumin concentrations were determined using standard analytical methods.
Statistical evaluation
Total use of morphine for each trial day was calculated by adding scheduled morphine doses and rescue morphine consumption. Samples obtained less than two hour after the administration of a morphine rescue dose were excluded from the analyses.
Day-to-day variations of morphine and its metabolites are presented as biological coeffecients of variation. This biological variation (CVBiological), expressed in terms of percent coefficient of variation, was calculated for each patient in each trial period using the equation [15, 16]:
CV
Biological
= CV
Observed
- CV
Analytical
The observed coefficients of variation (CVObserved) for morphine, M6G and M3G, which represent the variation in serum concentrations for each patient during each trial period, were calculated using the equation [17, 18]:
At least three observations were needed in order to calculate an observed coefficient of variation.
Statistical comparisons between the trial days and trial periods were performed using one-way analysis of variance tests. Due to multiple comparisons statistical significance was defined as p < 0,01.
The statistical software SPSS version 9.0 for Windows was used throughout the analyses.