Diclofenac readily penetrates the cerebrospinal fluid in children

Abstract

AIMS

The primary aim was to study the cerebrospinal fluid (CSF) penetration of intravenous diclofenac in children. The secondary aim was to evaluate the plasma diclofenac concentration at the onset of wound pain after inguinal surgery in children.

METHODS

A total of 31 children (24 boys) aged 3 months to 12 years received a single intravenous injection of diclofenac 1 mg kg⁻¹. Paired CSF and blood samples were obtained 5 min to 22 h (median 69 min) later. In children having inguinal surgery a second blood sample was obtained at the time that the children felt wound pain for the first time after surgery. Diclofenac concentrations in CSF, plasma and protein free plasma were measured by gas chromatography with mass spectrometric detection.

RESULTS

In the 28 CSF samples obtained at 5 min to 3 h 43 min after injection, diclofenac concentrations ranged between 0.5 and 4.7 μg l⁻¹. At 5.5 h the CSF concentration was 0.1 μg l⁻¹, and no diclofenac was detected in the two CSF samples obtained at 22 h. The median of plasma diclofenac concentration at the time when pain returned after inguinal surgery was 104 μg l⁻¹ (range 70–272 μg l⁻¹). No serious or unexpected adverse effects were reported.

CONCLUSIONS

Diclofenac penetrates the CSF rapidly, and a sufficient concentration to inhibit cyclooxygenase enzymes is sustained for up to 4 h.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• Diclofenac, a nonselective nonsteroidal anti-inflammatory drug,, exerts analgesic action both in the peripheral tissues and in the central nervous system by inhibiting cyclooxygenase enzymes COX-1/2, but central nervous system penetration of diclofenac has not been evaluated in humans.

WHAT THIS STUDY ADDS

• Diclofenac penetrates the cerebrospinal fluid rapidly, and after a single intravenous dose of 1 mg kg⁻¹, sufficient concentrations to inhibit COX-1/2 are sustained for up to 4 h.

Introduction

Diclofenac is the most commonly used intraoperative nonsteroidal anti-inflammatory analgesic (NSAID) in children in the UK and Ireland [1]. In the treatment of mild and moderate acute pain diclofenac is sufficient as a sole agent, and in severe pain it has an opioid-sparing effect [2, 3]. Some studies have evaluated the pharmacokinetics of diclofenac in paediatric patients [4–6], but to our knowledge its central nervous system (CNS) penetration and analgesic concentration has not been described in children.

The CNS pharmacokinetics of NSAIDs is of interest because animal studies show that these compounds elicit analgesic action not only in the peripheral tissues but also in the CNS [7]. An estimate of analgesic concentration may help to plan an optimal dose regimen for diclofenac in paediatric patients with acute pain. Therefore, we designed the present study, in which the primary aim was to evaluate the cerebrospinal fluid (CSF) penetration after intravenous injection of diclofenac 1 mg kg⁻¹ in children. A paired blood sample was obtained close to the time of CSF sampling to measure whether there was a correlation of CSF concentration to total and protein free plasma diclofenac concentrations. The secondary aim was to evaluate the plasma diclofenac concentration at the time the children felt wound pain for the first time after inguinal surgery with levobupivacaine spinal anaesthesia.

Methods

Subjects and sampling

The study was conducted in Kuopio University Hospital (Kuopio, Finland). It was recorded in the EudraCT database (no. 2004-001702-27) and the Finnish National Agency for Medicines was notified (no. 161/2004). The study was conducted in accordance with the Declaration of Helsinki. The Research Ethics Committee of the Hospital District of Northern Savo, Kuopio, Finland (no. 120/2004) approved the study. Informed written consent was obtained from all parents and children gave their assent if old enough.

This study is part of our research project ‘Non-opioid-analgesics in CSF in children’ and the protocol has been described in detail earlier [8]. Briefly, children with no or mild systemic disturbance (the American Society of Anesthesiologists physical status 1 or 2) aged 3 months to 12 years who were scheduled for surgery in the lower part of the body with spinal anaesthesia were eligible for the study. Children with any concomitant disease, with a body mass index percentile equal to or greater than the 95th for age and gender, as well as those with any contraindications to diclofenac or spinal anaesthesia, were excluded.

The children received a single 10-min intravenous injection of diclofenac 1 mg kg⁻¹ (Voltaren 25 mg ml⁻¹; Lot no. S0258, BB 05 2007; Novartis Finland Oy, Espoo, Finland), diluted in 20 ml 0.9% saline buffered with 0.5 ml NaHCO₃ 7.5%. The injection was given 5 min to 22 h (median 69 min) before lumbar puncture for spinal anaesthesia. The children received a standard transmucosal midazolam-ketamine premedication and were sedated with midazolam and thiopental to ease lumbar puncture. One millilitre of CSF was collected into a polypropylene tube during lumbar puncture before the injection of local anaesthetic. Within 5 min, an indwelling catheter was inserted in a dorsal foot vein and a 3-ml blood sample was obtained.

After surgery in the post-anaesthesia care unit, vital signs, adverse effects and pain were monitored by trained research nurses and one of the researchers (H.K., E.K., M.La.). In eight children with inguinal surgery, pain was assessed every 15 min at rest and with light pressure (20 N) on the wound area. On the first occasion that the child expressed or was assessed as having wound pain, a second blood sample was obtained for the determination of plasma diclofenac concentration at the onset of pain. Thereafter the child received paracetamol 15 mg kg⁻¹ and ketoprofen 1 mg kg⁻¹ intravenously. For rescue analgesia children in pain (pain score >3 at rest and/or with >5 with 20 N compression in a numeric rating pain scale 0–10) were provided fentanyl 1 μg kg⁻¹ (n = 5) or oxycodone 0.05 mg kg⁻¹ (n = 4).

Diclofenac assay

Blood samples were collected into heparinized tubes, plasma was obtained by centrifugation at 3000 g at 20°C for 10 min, and the samples were stored at −80°C in polypropylene tubes and protected from light.

Diclofenac concentrations were measured in CSF, plasma and protein free plasma samples by gas chromatography with mass spectrometric detection. A modified method previously described by Mannila et al.[8] was used for sample pretreatment: briefly, CSF (1 ml), plasma (100 μl) or protein free plasma (100 μl) sample containing the internal standards (ketoprofen 251 ng per sample was internal standard I and naproxen 45 ng per sample was internal standard II; Sigma-Aldrich, Steinheim, Germany). Sample was acidified with 20 μl of 1 m HCl and applied into solid-phase extraction cartridges (Discovery, DSC-18, 1 ml 100 mg⁻¹; Supelco, Bellefonte, PA, USA). Cartridges were washed with water (1 ml), and compounds of interest were eluted with ethyl acetate (1 ml). Extracted samples were evaporated to dryness under nitrogen and 300 μl of 4.0% (v/v) pentafluorobenzylbromide in toluene was added in a screw-capped glass test tube. To the sample 50 μl of triethylamide was added. The tubes were vortex mixed and heated at 110°C for 2 h. After cooling samples to room temperature, samples were evaporated to dryness under nitrogen and extracted with acidified water and toluene. The toluene phase was analysed.

Nonprotein-bound drug samples were obtained by ultrafiltration of 300 μl of plasma using Centrifree® Micropartition Devices (Millipore Corp., Bedford, MA, USA). Samples were centrifuged for 15 min at 1500 g at 22°C. A 150-μl aliquot of filtrate was stored at 4°C until assayed.

Diclofenac, ketoprofen (internal standard I) and naproxen (internal standard II) were identified and quantified by Agilent Technologies GC-MS system (Agilent Technologies, Palo Alto, CA, USA). The system consisted of a 7683 autosampler and a 6890 N gas chromatograph coupled to a 5973 N mass spectrometry. Data were collected using Agilent Technologies Enhanced ChemStation Software (Version C.00.01.08). Chromatographic conditions were as follows: a 30 × 0.25 mm i.d., cross-linked 5% phenyl methyl siloxane capillary column with 0.25 μm film thickness (HP-5MS; Agilent Technologies) was used with splitless injection. The injection port temperature was 250°C and injection volume was 1 μl. The temperature program was as follows: from 110°C to 280°C at 40°C min⁻¹ and hold at 280°C for 5 min. The carrier gas was helium with constant flow of 37 cm s⁻¹. The temperatures of the mass spectrometer detector transfer line heater, ionization source and quadrupole were maintained at 290°C, 150°C and 150°C, respectively. The mass spectrometer (MS) was operated in negative chemical ionization mode (electron energy 220 eV and emission current 69 μA) and the reagent gas used was methane. The MS was operated in the selected ion monitoring mode (25 ms dwell time); the masses followed were 294, 253 and 229 atomic mass units to pentafluorobenzyl derivative of diclofenac, ketoprofen and naproxen, respectively.

CSF and plasma ultrafiltrate methods were linear over the concentrations of 0.1–23 ng per sample, and plasma method over the concentrations of 5.8–2300 ng per sample. The intraday precision of the method was determined by analysing quality control samples at the three different concentrations. Concentration levels were 0.5, 9 and 18 ng per sample for the determination of CSF and protein free plasma samples giving the intraday precision CV% values of 8.3, 5.2 and 6.5%, respectively. The interday precision determined with quality control samples at the above-mentioned concentrations gave CV% values of 11, 4.7 and 5.4%, respectively. The accuracy determined with quality control samples at concentrations 0.5, 9 and 18 ng per sample was 113, 92 and 90%, respectively. Concentration levels were 93, 467 and 1869 ng per sample for the determination of quality control plasma samples giving the CV% of 0.2, 0.9 and 0.4%, respectively. The interday precision determined with quality control samples at the above-mentioned concentrations gave CV% values of 0.4, 1.4 and 8.9%, respectively. The accuracy determined with quality control samples at concentrations 93, 467 and 1869 ng per sample was 95, 103 and 113%, respectively. The recovery of the method ranged between 70 and 80%. Validity of the ultrafiltration technique was studied with drug-free human plasma, which was spiked with diclofenac to give a final concentration of 2000 ng ml⁻¹. After ultrafiltration the samples (100 μl) were analysed with a GC-MS. Nonprotein-bound plasma concentration was 1.7 ng ml⁻¹ corresponding to 0.08% from the total plasma concentration with a precision of 4.9% (CV%).

Statistics

No formal sample size calculation was performed, but a sample of 30 children was considered to provide sufficient information on CSF penetration of diclofenac in children. Data were entered and analysed with the Statistical Package for Social Sciences (SPSS software version 14.0 for Windows; SPSS Inc., Chicago, IL, USA). Because there was no control group, descriptive results (number of cases or median with range) are presented. Correlations between diclofenac concentrations and patient characteristics were tested with the Pearson correlation test and independent samples t-test. A P-value of 0.05 was considered as the limit of statistical significance.

Results

Written consent was obtained from the parents of 32 children, with no refusals. One child was excluded from the study because the operation was postponed due to administrative reasons, leaving 31 children in the study group. All children received diclofenac and the samples were collected as described in the protocol, with no protocol violations. Patient characteristics are presented in Table 1.

Table 1.

Patient characteristics

Variable	n = 31
Gender (male/female)	24/7
Age (years and months)
Median	4 years 8 months
Range	3 months to 12 years 9 months
Weight (kg)
Median	21
Range	6–60
Height (cm)
Median	111
Range	60–169
Operations (n)
• Inguinal surgery	11
• Urological surgery	10
• Orthopaedic surgery	5
• Other	5

Data are number of cases or median with minimum–maximum.

Diclofenac was detected in 29 of the 31 CSF samples. CSF diclofenac concentrations ranged between 0.5 and 4.7 μg l⁻¹ in the 28 samples obtained at 5 min to 3 h 43 min after dosing. In the sample obtained at 5 h 30 min the CSF diclofenac was 0.1 μg l⁻¹, and no diclofenac was detected in the two CSF samples obtained at 22 h after drug administration. There was a negative correlation between CSF diclofenac concentrations and children's age, height and weight (r = −0.47, P = 0.01; r = −0.56, P = 0.001; and r = −0.44, P = 0.017, respectively). The CSF/protein-free plasma concentration ratios ranged between 0.2 and 5.2 (2.2), and the CSF/total plasma concentration ratios between 0.0002 and 0.048 (0.005). There was no correlation between CSF and protein-free or total plasma concentrations of diclofenac. CSF and plasma diclofenac concentrations are presented in Table 2 and Figure 1.

Table 2.

Cerebrospinal fluid and plasma diclofenac concentrations in each patient after i.v. injection of 1 mg kg⁻¹ (n = 31)

Sampling time (h and min)	CSF concentration (μg l−1)	Unbound plasma concentration (μg l−1)	Total plasma concentration (μg l−1)	Ratio CSF/unbound plasma concentration	Age (years and months)	Height (cm)	Weight (kg)
5 min	0.7	1.2	2073	0.60	6 years 11 months	120	22
6 min	0.6	1.0	1005	0.66	4 years 3 months	113	21
10 min	3.6	1.6	1914	2.3	3 months	60	7
10 min	0.9	2.0	4232	0.46	7 years 6 months	111	19
11 min	0.8	1.9	2157	0.42	2 years 11 months	96	14
14 min	0.8	3.8	1618	0.20	4 years 8 months	111	24
17 min	0.5	0.9	1308	0.58	6 years 2 months	120	23
20 min	1.2	1.2	2024	1.0	2 years 2 months	90	14
20 min	1.2	1.2	2213	0.95	12 years 5 months	169	60
24 min	2.4	1.0	1442	2.3	1 year 5 months	83	11
30 min	1.7	0.8	718	2.3	3 years 11 months	107	21
43 min	3.1	0.6	666	5.2	1 year 5 months	80	11
44 min	2.2	0.7	540	3.0	6 years 4 months	116	22
51 min	1.6	0.8	359	2.2	12 years 9 months	141	34
60 min	4.7	0.9	753	5.1	5 months	62	6
1 h 9 min	1.4	0.6	276	2.4	7 years 11 months	131	32
1 h 23 min	1.4	0.4	138	3.4	2 years 6 months	85	11
1 h 28 min	1.7	0.8	253	2.1	10 years 1 month	143	31
1 h 54 min	1.4	0.4	591	3.7	4 years 11 months	114	20
2 h 2 min	1.3	ND	111		5 years 7 months	118	22
2 h 4 min	1.1	ND	148		6 years	130	29
2 h 19 min	1.1	ND	113		1 year 4 months	75	11
2 h 25 min	0.9	0.4	116	2.3	9 months	76	11
2 h 28 min	0.9	0.5	77	1.9	7 years 10 months	122	25
2 h 50 min	3.2	ND	66		10 months	75	10
3 h 10 min	0.7	ND	55		3 years 5 months	97	14
3 h 11 min	1.0	0.4	91	2.4	10 years 8 months	140	41
3 h 43 min	0.6	ND	80		6 years	122	31
5 h 33 min	0.1	14.8	57	0.008	11 years 1 month	136	32
21 h 56 min	ND	ND	ND		9 months	68	8
22 h	ND	ND	ND		3 months	67	9

ND, not detected.

Figure 1.

Cerebrospinal fluid and plasma diclofenac concentrations after intravenous injection of 1 mg kg⁻¹ in children (n = 29). CFS, (□); Plasma, (♦)

Diclofenac was detected in 29 of the 31 plasma samples obtained after the lumbar puncture; the concentrations ranged between 55 and 4232 μg l⁻¹. Nonprotein-bound diclofenac concentrations ranged between 0.4 and 3.8 μg l⁻¹ in 22 of the 31 samples. One sample was an outlier with nonprotein-bound diclofenac 14.8 μg l⁻¹. Diclofenac was 99.4–99.9% (99.9%) protein bound in plasma. Diclofenac plasma concentrations did not correlate with children's age, height, weight or gender, and they were comparable to those reported by Korpela and Olkkola [4].

A second blood sample was collected from eight children who had undergone herniotomy or orchidopexy. These children developed pain 1 h 37 min to 5 h 16 min (3 h 3 min) after diclofenac injection, 1 h 31 min to 4 h 56 min (2 h 14 min) after spinal anaesthesia and 20 min to 4 h 32 min (1 h 38 min) after the end of surgery. The diclofenac concentrations at the onset of wound pain ranged between 70 and 272 (104) μg l⁻¹.

Seven children experienced seven nonserious adverse effects: agitation (n = 1), nausea (n = 2), vomiting (n = 2) and shivering (n = 2).

Discussion

Diclofenac is a commonly used intraoperative NSAID known to have analgesic action both in the peripheral tissues and in the CNS. However, little is known about diclofenac CSF penetration in humans. Zecca et al.[9] have reported two adult patients with CSF concentrations of 3.5 and 8.5 μg l⁻¹ at 2 and 12 h after intramuscular diclofenac 75 mg. To our knowledge, this is the first study describing CSF penetration of diclofenac in children. The present study indicates that in children diclofenac readily penetrates CSF after an intravenous dose of 1 mg kg⁻¹; diclofenac was detected in the earliest samples collected at 5 and 6 min after the injection and significant CSF concentrations of ≥0.5 μg l⁻¹ were sustained up to 4 h after dosing. In the three samples obtained at 6 h or later after administration, CSF diclofenac was below the range of the assay, 0.4 μg l⁻¹.

It has been shown that spinal cord cyclooxygenase (COX)-1 plays an important role in the first nociceptive component of acute, postoperative pain, and that inducible spinal cord COX-2 is involved in the later inflammatory primary hyperalgesia [10]. Thus, diclofenac, which blocks COX-1 and COX-2 equipotently [11, 12], should be an appropriate non-opioid analgesic for the management of acute postoperative pain. Furthermore, in experimental studies in rats it has been shown that diclofenac dose dependently reduces c-Fos expression in the superficial and deep laminae of dorsal horn at L4–L5 segments [13]. Intrathecal diclofenac has a 100-fold antinociceptive potency when compared with systemic diclofenac [7]. Thus, we believe that the lumbar CSF concentrations observed in the present study of approximately 1% of concurrent plasma concentration should be sufficient to exert analgesic action in the CNS.

COX inhibition has been supposed to be the fundamental mechanism of NSAID analgesic action. However, COX-independent mechanisms, including peroxisome proliferator-activated receptor-gamma (PPARγ) activation may also be involved [14]. PPARs are important regulators of energy homeostasis and inflammation in the body, and are present in the brain and spinal cord [15]. Indomethacin and ibuprofen activate PPARγ, but only at high concentrations [14]. Diclofenac is unique among NSAIDs, as it interacts with PPARγ signalling at clinically relevant concentrations [16]. It has been argued that the PPARγ activation in conjunction with inhibition of prostaglandin synthesis may synergistically augment the effects of diclofenac on spinal nociceptive processing. PPARγ activation also suppresses the activation of microglia, suggesting that diclofenac may be of benefit in chronic neuroinflammatory conditions such as persistent postsurgical pain [17]. Therefore, we believe that the action of diclofenac in the CNS may include COX inhibition and anti-inflammatory actions via PPARγ signalling at the concentrations achieved in the present study.

The secondary aim of the study was to evaluate plasma diclofenac concentrations at the time that the children felt wound pain for the first time after inguinal surgery. The median plasma diclofenac concentration at the onset of wound pain was 104 μg l⁻¹ and that concentration was sustained for 2–3 h after intravenous diclofenac 1 mg kg⁻¹. As far as we are aware, this is the first attempt to estimate the analgesic plasma concentrations of diclofenac in children. However, one of the main limitations of the present study is that the decrease in diclofenac concentrations occurred at the same time as the wearing off of the anaesthesia, and thus these results would not allow any conclusions to be drawn.

The optimal dose of diclofenac in children has not been established [6, 18, 19]. The maximum daily dose of diclofenac for the management of pain is 2 mg kg⁻¹, but in rheumatic disorder a dosage of 3–5 mg kg⁻¹ divided into two or three doses has been approved [20]. We suggest that the larger dose would also be more appropriate in the management of acute postoperative pain. The present study with 1 mg kg⁻¹ of diclofenac indicates that in clinical use either a higher initial dose or repeating the dosage after 3–4 h should be used in order to maintain sufficient analgesic concentrations. In a recent study, children having major surgery were administered diclofenac 1.5 mg kg⁻¹ intravenously followed by 2 mg kg⁻¹ per rectum b.i.d. With that dosage, plasma trough concentrations were >100 μg l⁻¹ in all children and the need for opioid analgesic was 50–70% less than in the control group without any background analgesia [21].

In paediatric pharmacotherapy the formulation may be more important than the compound used. In the present study diclofenac was administered intravenously. Intravenous dosing ensures that all drug enters the circulation. However, unlike in Finland and in the UK, parenteral diclofenac is not approved for intravenous use in all countries. Diclofenac injection is approved also for intramuscular use, but we do not recommend it: first, because children dislike intramuscular injections [2], and second, because diclofenac is associated with local irritation and pain, and rare cases of serious injection site necrosis (Nicolau's syndrome) have been reported [22].

The two enteral dosage forms of diclofenac are suppositories and tablets. Suppositories can be used during anaesthesia, but awake children dislike suppositories [2]. Dosing by mouth can not be used in the immediate postoperative phase, but later, when gastrointestinal function has recovered, tablets may be used. Our recommended clinical practice for the management of acute pain in children is to give the first doses of analgesics intravenously and then continue treatment with tablets.

In summary, the rate and extent of diclofenac penetration into the CSF is sufficient to provide COX-1/2 inhibition up to 4 h after an intravenous dose of 1 mg kg⁻¹.

Competing interests

None to declare.