Therapeutic drug monitoring (TDM) is defined as: “measuring serum concentrations of a drug in a single or multiple time points in a biological matrix after a dosage” (1). TDM is useful when the relationship between drug plasma concentration and effect, is stronger than between drug dosage and effect and is vital when measuring drugs with a narrow therapeutic range (only small differences between the therapeutic and toxic effects). The therapeutic range is a concept meaning that within this range, most patients being treated with the particular drug with have clinical benefit, with limited side effects (1) (2). Furthermore, it aims to improve clinical outcomes if toxicity is suspected; when clinical effects are not easily measurable; if sub-therapeutic concentrations are suspected; if compliance is questionable and if concentrations resulting from a given dose are unpredictable due to high inter/ intra patient variability (3).Knowing the optimal dose for a patient can be difficult in conditions such as epilepsy, with clinical signs of toxicity not always apparent and therapy failing in 1/3 of patients (4) (5).
Plasma TDM has a limited role and is not an absolute value, as everyone responds to treatments differently and problems arise predicting plasma concentrations from the dose given because many pharmacodynamics and pharmacokinetic factors such as absorption, distribution, metabolism and excretion lead to individual variability (2). This is especially true in infants, who for example, have more rapid metabolism than adults (2) (6), lower protein binding which increases the free fraction of drugs and a less mature glomerular filtration rate (GFR) (7) which means drugs may be retained for longer in the those <6 months old. It is not surprising that validated dosage regimens are limited in this population, with approximately 70% of drugs prescribed to children and >93% prescribed to critically ill neonates being done so off label (8) (9).
The use of serum/plasma in TDM however, is well established and the most clinically used method (10) (11) having been used since the 1960s (12), nevertheless extensive blood sampling in infants is not practical, or ethical due to pain, anxiety and risk of infection (13). There are a significant number of validated matrices available for TDM in adults’ however, these do not necessarily apply to infants. The matrices available for adults, but inappropriate for infants include tears, sweat and hair due to lack of availability in infants, exhaled breath due to the lack of variability of drugs available for TDM and cerebrospinal fluid due to pain concerns. TDM in infants is possible with careful selection of the most appropriate method, which is in line with strict guidelines on patient safety (14). To be ideal it must also be simple, vigorous, fully validated and inexpensive. Table 1 reflects the available matrices for TDM in infants and discusses the advantages and disadvantages of these.
Technological advances in TDM in the past four decades have led to saliva testing becoming more popular in the diagnosis of diseases, disease progression and detection of drugs (15) (16). It has become a valuable clinical method of detection due to its non-invasive nature, no need for specialist training or equipment, patient preference and the ability to compare easily with plasma concentrations (16) (17). This is because most compounds found in blood are also present in saliva. In infants this method seems ideal to prevent distress and ensure patient safety (18). If there is a consistent comparison between saliva and plasma, and if the biological response to the drug is proportionate to its plasma concentration, then salivary concentrations will provide a valuable measurement of TDM.
Many studies have been conducted in adults in regards to saliva TDM, with it being a proven accurate method in relation to many drugs including artemisinin (20), digoxin (21), lamotrigine (22) and a number of other antiepileptic drugs (23). However, these reference ranges do not automatically transfer to infants, and the accuracy of salivary TDM in infants is still an area of debate (24) (25), with confusion between collection methods and stimulated or unstimulated saliva.
Research involving paediatrics has been prioritised recently to increase the evidence in paediatric clinical pharmacology, without this, estimates are used which are not always reliable. In spite of this, the number of studies involving vulnerable infants are still low due to poor consent rates, limited blood availability, low volume drug concentration assays and a lack of expertise in this area (26). Studies in paediatrics are usually disease orientated and expensive as drug companies are unable to recover costs in this limited population (27). Consequently, there is a need for novel and accurate study designs which decrease these barriers, saliva TDM fills this gap and reduces these barriers.
This review aimsto explore the accuracy, applicability and advantages of using saliva TDM in infants <24 months.
Saliva as a biological matrix
Saliva or oral fluid is a clear and slightly acidic (pH 6.0-7.0) complex fluid which contains about 97% water, followed by electrolytes, immunoglobulins, enzymes and proteins (28) and is produced by the major and minor salivary glands, oral mucosa and gingiva (29) (15). Saliva has many important roles in the functioning of the body such as oral digestion and tissue lubrication (30). Furthermore, neonates have a higher salivary secretion pattern than adults (31). Blood sampling in premature infants is difficult due to low blood volumes, and saliva sampling would therefore be ideal for this patient group.
As a biological matrix for TDM, saliva has many advantages, including being inexpensive (18), non-invasive, rapid and with less risk of needle stick injuries and with the potential for collection to be performed at home (32). There is also the potential for samples to be sent directly to the laboratory for analysis (33) allowing round the clock sampling, which assists in the identification of patients with small daily fluctuations who may require only minimal or no fractionation of the daily dose (34) (35). There is minimal skill needed by the researcher and no effort from patients; therefore, it is preferred by patients and parents and may increase both compliance and patient-clinician relationship (18). In infants a major advantage is averting blood volume loss and reducing risk of trauma and infection from venous punctures (36).
However, a number of disadvantages for this type of sampling also exist- it provides a small sample size, especially in infants, which is viscous. Interference with food, drinks and oral medication may cause errors in the analysis (20). There is a lack of standardisation in procedures which may explain some of the conflicting results seen in studies. Another disadvantage is that saliva is not normally collected in routine procedures, whereas blood often is. Furthermore, in patients with severe mucositis or other oral mucosa/salivary gland damage and in patients where oral fluid production is reduced it may not yield enough sample volume. Some children and adults have also suggested that they do not like the taste of citric-acid and found the idea of spitting unfavourable (18) (35).
Factors affecting drug concentration correlation in saliva
Many of the constituents of this complex matrix pass though cells via transcellular, passive intracellular diffusion and active transport, or paracellular routes by extracellular ultra-filtration within the salivary glands or through gingival sulcus (37). Drug passage into saliva follows the general principles of movement across biological membranes (38). The most common mechanism being passive diffusion, which is limited to non-protein bound and non-ionized molecules with molecular weights less than 500 Da. The unbound form is the pharmacologically active component, whereas the bound is the reserve which helps maintain the steady state during the process of administration, absorption, distribution, metabolism and excretion (39). Very little of an administered drug is bound to protein in saliva. Binding of individual drugs to plasma protein differs little between subjects where liver and kidney function are normal (although it is known to reduce with temperature rise) (40).
The saliva/plasma concentration ratio (S/P) is a very useful parameter to know when assessing drugs contained within the saliva. This ratio is determined by a number of physiological factors which affect the passage of drug from blood to saliva. These include the pH of the oral fluid and plasma, the saliva flow rate and the pathophysiology of the oral cavity (40). With the exception of salivary pH, most of these remain constant. The primary drug properties which also play a key role in the passage of drugs into saliva include the pKa of the drug, its molecular weight, spatial configuration and lipid solubility of the analyte and degree of protein binding in plasma and saliva (15). Figure 1A illustrates the primary properties of the drug that influence penetration into saliva with 1B displaying the physiological properties of the saliva that influence drug penetration.
Plasma pH varies within a very narrow range and so does not affect S/P ratio significantly. However, as only non-ionized drugs can cross biological membranes and the pH of saliva is usually less than that of plasma, basic drugs usually concentrate in oral fluid. This phenomenon is termed ion trapping because after the free drug crosses into oral fluid it may ionise and become ‘trapped’ in the more acidic fluid. In contrast to this, acidic drugs may concentrate in plasma. It is also suggested that although weak acids and bases may not be affected by salivary pH, stronger one’s ratios could alter by as much as 100-fold. For Acidic Drugs (pKa>8.5) or basic drugs (pKa<5.5), S/P unbound ratio remains constant irrespective of changes in saliva pH, as a result of changes in salivary flow (41). Weak bases are thought to be the least reliable compounds to monitor in saliva as they are sensitive to pH changes, while neutral compounds are the most reliable as they probably reflect the unbound drug in serum.
Saliva flow rates vary significantly between individuals and conditions (42). Its flow is regulated by neurotransmitter and hormone release and may be affected by resting, healthy or stimulated states. A rapid flow rate has been shown to increase linearity of S/P ratio (43), while some drugs can actually affect the flow (44). Stimulated states can influence the normal oral pH and also sodium content of saliva increases linearly with stimulation (37) (45) (46). Furthermore, clinical condition is also thought to affect salivary flow, especially in critically ill patients (47). This salivary flow rate can influence the S/P ratio by increasing the pH of oral fluid if stimulated. For example, phenytoin, a weak acid, decreases in concentration in oral fluid as salivary flow increases. Salivary flow has a limited effect on the S/P ratio of other anti-epileptic drugs however (42). Compounds with pKa values < 5.5 or >8.5 are not significantly influenced by small changes in salivary pH due to flow rate changes; S/P ratios tend to remain relatively constant for these analytes. For neutral, non-protein bound drugs, e.g. ethanol, the S/P ratio is 1, and these drugs which are never ionised, regardless of pH, will not be influenced by alterations in saliva flow rate. Another aspect of salivary flow that may influence the S/P ratio is dilution of analyte concentration due to the increased volume of saliva following stimulation.
Stimulation of saliva can be conducted in a number of ways, chewing paraffin wax, parafilm, teflon, or using chemical stimulus, namely citric-acid (48). Stimulated saliva has a number of benefits compared to resting saliva as it produces a better flow rate, greater sample size, the pH gradient between plasma and saliva is smaller, the variability in the S/P ratio has shown to be lower in some drugs and it produces better specimens (less viscous) for analysis (49). However, it must be used adequately, as too much may interfere with assays and some substances including glucose, proteins and drugs (29) (36) (50) (51). It has also been noted that highly lipophilic molecules may be absorbed by parafilm (52).
Saliva Collection methods
There have been many collection methods tried for saliva, and although spitting into a container seems to be the easiest method in adults and children, this is not practical for infants. When considering infants, it is thought that wiping the mouth with sterile gauze and rinsing with 5ml of sterile water and waiting for 5 minutes helps to reduce the risk of contamination (53) from either oral medication or mothers milk. Some research has shown that by turning the infant onto their side to let the saliva pool in the cheek was the easiest method of collection, whereas other authors have gotten the saliva from under the tongue and have collected it using various devices such as pipettes (54), mucous extractors (38), adapted pacifiers (55), suction devices (24) (56) and more recently salivettes/salimetrics collection devices (57).
These devices have all gotten advantages/disadvantages; Pipettes have limitations due to saliva viscosity, the adapted pacifier is not applicable to all infants and suction devices are more troublesome than the other methods. The Salivette collection device has been specifically designed for the purpose and consists of a cotton swab which is placed in the mouth until saturated, it is then placed it its container where the fluid can be collected. It has shown to achieve good recovery for most drugs, it is practical, convenient and easy to use. Of course, there are downsides to even this collection method as some drugs may absorb onto the cotton swab (58). Despite this, it remains suitable for most drugs.
The type of collecting tube also needs to be taken into account when beginning a saliva collection. Studies have shown that tubes containing serum separator gels can significantly affect the determination of some drugs such as phenytoin (59) (60) or ribavirin (61). Therefore, it is strongly recommended to evaluate the matrix effect of collecting tubes (3).
Current drugs assessed for TDM in infants
The review question was defined using PICO as follows: P: Infants (<24 months) being treated with any medication, I: Salivary TDM vs C: Traditional methods and O: accuracy, advantages/disadvantages and applicability to practice. The following databases were searched; MEDLINE, CINAHL, EMBASE, Scopus, Google Scholar, The Cochrane Library and PubMed up to and including August 2016. 24 published articles were eligible for full review and in all cases the primary sources were used unless no subsequent peer reviewed article was published.
Thirteen different compounds were considered to be suitable for inclusion within this review. The range of therapeutic indications included analgesic, antimicrobial, apnoea, arrhythmias, asthma, and epilepsy. Of the compounds chosen, they were categorised as being acidic (n=4), basic (n=3) or neutral (n=6). In the case of an amphoteric compound, such as morphine, it was included within the category to which it was strongest, in this case basic. Acidic compounds were deemed to have a pKa less than 10, and basic compounds were considered as having a pKa greater than 4.0. Any pKa value outside of this range (whether acidic or basic) would not be substantially affected by the physiological pH range of saliva (≤ 0.1% ionised) and was thereby categorised as neutral. Table 2 displays the compounds assessed within this review and includes the chemical structure as well as both the physicochemical and pharmacokinetic parameters.
As the various authors have described the correlation between the experimental blood plasma levels of a compound and that of saliva levels in different manners, for the purpose of this review, the R value has been categorised as either excellent, very good, good, fair or poor. The parameters of each can be seen in Table 3 with Table 4 summarising the information on each of the thirteen compounds.
Acidic compounds (pKa ≤ 10)
Four acidic compounds have previously been studied for saliva content in infants (table 2) with the pKa ranging from digoxin (7.15) to theophylline (8.81). The individual compounds are discussed below.
In neonatal intensive care units, theophylline is one of the most regularly drugs used for apnea of prematurity (62), however it has potentially serious side effects with infants having a decreased capacity to metabolise the drug, which then alters with age. Many factors such as concomitant drug use can affect theophylline pharmacokinetics and a high interindividual variability means TDM is essential. It is a weakly acidic compound with pKa of 8.81, meaning that a variation in saliva pH from 6 – 7 will have a limited effect on the % ionisation (from 0.15 to 1.5%).
In a study by Toback et al (1983) (63) 8 premature infants had blood (0.25ml) collected by heel prick and unstimulated saliva (0.25-0.50ml) collected by suctioning with a syringe 4 and 8 hours after IV aminophylline administration. Since no stimulation was used, collection took 30-60 minutes. Samples were centrifuged and stored at 4oC for 48 hours or analysed immediately by HPLC. Excellent correlation was seen in this study (R=0.98 p<0.01) at both sample time points. The S/P ratio was 1.02 +/-0.09 and the coefficient of variance was 8.9% (95% confidence). No limit of quantification was mentioned. They suggested that serum and saliva concentrations of theophylline are approximately equal and that it is a valid method of TDM in premature infants if serum is unavailable.
In a later study by Siegel et al (1990) (64), saliva and citric acid stimulated saliva were collected along with blood from 31 children (2-17.5 years) receiving oral aminophylline. Saliva was collected by spitting/suction at least 7 hours after the previous dose and was centrifuged and stored at 4oC until analysis by an enzyme-multiplied immune assay technique (EMIT). There was a significant correlation (p<0.001) between total serum or unbound serum, and either stimulated or unstimulated saliva. The correlation was highest (R=0.98) when total serum and stimulated saliva were compared, and lowest (R=0.89) with unbound serum vs unstimulated, however total serum vs unstimulated had an excellent correlation also (R=0.97). The lower correlations with unstimulated saliva was thought to be due to greater variability in the theophylline concentration in the saliva due to flow rates, pH etc. This contradicts other studies which believed theophylline concentration in saliva was not affected by these factors (65) (66) as higher theophylline concentrations were seen in unstimulated saliva. Overall they believed stimulated saliva sampling would be significant for TDM, with again, treating levels between 9-11µg/ml as toxic and conducting blood sampling.
In a larger study conducted in 2001 by Culea et al, (62) salivary TDM in 40 infants was deemed an adequate and reliable method for theophylline. Only 13 of the infants were <24 months and so the data reviewed only looked at group B of the study. Blood and stimulated saliva was collected from the IV group between 2 and 8 hours after dosing. Saliva (0.5-1ml) was collected by aspiration and no storage methods were noted. Analysis was conducted using GC-MS. A good correlation was seen between serum and salivary theophylline concentrations (R=0.97 p not stated). Their S/P ratio was 0.69 +/- 0.13.
In a more recent study by Chereches-Panta et al, (67) 13 infants <38 weeks’ old were studied. Unknown quantities of blood and stimulated saliva were collected by aspiration, 2 and 4 hours following the dose. Samples were stored at -20oC for 48-72 hours before analysis using Mass Spectrometry (MS). Excellent correlation was seen (R=0.973 p<0.01), with no significant difference between the timings of sample collection. The S/P Ratio was 1.53 +/-0.27. In this study they also looked at the post conception age of the children and determined that there was a significant difference between the ratios in the two age groups (32.8 +/-2 weeks/ 37.1+/=0.8 weeks). However, they concluded that the strong correlation and small interindividual variability would allow the use of saliva to monitor plasma concentrations, taking into consideration that higher concentrations may be seen in younger infants.
An earlier study by Khanna et al (1980) (24), looked at 10 premature infants receiving theophylline IV. Blood was obtained by heel prick (150µl), centrifuged and serum collected. Unstimulated saliva (50µl) was then collected 2-3 times 4-6 hours after drug administration, using a sterile rubber ear syringe by suctioning- this was the most convenient method they tried, ranging from 2-5 minutes. Both blood and saliva were frozen until analysis by HPLC but conditions were not specified. The correlation was poor in this study (R=0.70 +/-2.8). Only levels of 2µg/ml or higher were used due to errors when serum concentrations are extremely minute.Their S/P ratio of 1.24 is similar to the observations of others (68) (69) and they concluded that if salivary concentrations are above 8 µg/ml, serum concentrations need to be measured due to high deviations in saliva samples above this concentration. Based on this study, saliva may be utilised as an alternative to blood TDM for theophylline in premature infant’s dependent on the concentration, however high interpatient variability was a drawback to this studies results.
Saliva theophylline levels are only slightly higher than plasma unbound concentrations (70) (64) (71). With only small intra and interpatient variability (72) (73). This data shows that TDM of theophylline using stimulated or unstimulated saliva is an accurate method in infants when considerations are taken into account. Consideration needs to be given to the higher saliva/plasma theophylline concentration ratio in neonates as compared to children and adults. The conflicting serum vs saliva data regarding theophylline cannot be explained by salivary flow rate or pH fluctuations as the secretion of theophylline is not influenced by these factors (72), the method of citric-acid stimulation is the most suitable method of TDM for theophylline, however the high variability which has been seen is similar to that noted by others (74) (65). The difference seen between populations could be due to differences in protein binding in these groups, but this requires further exploration. It has, however, been suggested that neonates have minimal protein binding of theophylline (75) and that proteins in this group generally differ from that found in children and adults (76) (77). In addition, high plasma concentrations of free fatty acids, unconjugated bilirubin and steroid in neonates can compete with certain drugs for binding sites. It has also been suggested that pH changes in blood as little as 0.2 may affect the protein binding (78).
Prescribed for seizures, TDM of phenytoin is important due to its small therapeutic window, incomplete absorption, dose-dependent elimination kinetics and marked interpatient variability, including age-dependent metabolism (79). Phenytoin incurs a number of interactions, due to enzyme induction/inhibition, but also due to its high protein binding (90%) (80) meaning that displacement of a small amount of drug from the binding site may increase the free concentration significantly and therefore the pharmacologically active concentration (50)- this is important in combination therapy with valproic acid which is common practice for uncontrolled epilepsy. It is a weak acid (pKa 8.3) with 0.5% ionisation at pH 6, raising to 4.77% at pH 7.
Cook et al 1975 (81) conducted a study on 38 patients (8 months-62 years). Saliva (10ul) was generally collected by spiting/capillary-pipette-dilution system. Blood was collected simultaneously and was centrifuged and analysed by radioimmunoassay. No storage conditions were noted. The correlation was excellent (R=0.98 p<0.001) and the S/P ratio was 0.101 +/-0.003.
In a study by Goldsmith and Ouvrier (1981) (82) unstimulated saliva (1-2ml) and plasma was collected from 202 children (5 months-18 years). Both matrixes were centrifuged and supernatant collected for analysis and frozen in unspecified conditions. Saliva was collected using a mucous trap suction device. They studied two different assays methods, gas liquid chromatography (GLC) and EMIT. They found that there was better correlation with fresh samples compared to frozen ones. In the 15 infants (<24 months) a S/P ratio of 0.11 +/-0.03 was determined and there was no significant difference seen between the age groups. For the GLC assay the correlation was very good (R=0.94) and the EMIT was not as significant (R=0.86). The risk of contamination by oral ingestion was not considered until after the study was completed. in children theirs was 0.11 GLC and 0.12 EMIT which is in good correlation. No correlation was found between pH of saliva and S/P ratio (R=0.22).
In another study in 1981 by Mucklow et al (41), 17 infants (<24 months) were studied at 3 weeks, 6 weeks and at 3 month intervals for 1 year. Stimulated saliva was obtained by aspiration under the tongue using a mucous extractor and blood was collected on the final visit to the clinic. No drug had been taken within four hours of the sampling and storage conditions were not specified. The samples were analysed by GC. A highly significant correlation was demonstrated (R values not determined), with a mean S/P ratio of 8.4 +/-1.3. In adults and older children, it is established that for phenobarbitone a linear relationship exists between dosage and plasma concentration throughout the therapeutic range, whereas in phenytoin wide individual metabolism rates have been demonstrated. The relationship between dose and concentration is nonlinear so these results are not surprising for phenytoin.
In a study conducted by Lifschitz et al (1990) (34) 16 children and infants taking oral phenytoin were studied to determine the correlation between citric-acid stimulated saliva and blood. The samples were collected within an hour of each other using a syringe (1-2 ml). Water was rinsed around the infant’s mouth before the samples were taken. The samples were refrigerated for 24 hours before analysis by fluorescence polarisation immunoassay. The S/P ratio was 9.54 +/- 1.05 and the correlation between free/total phenytoin in plasma compared to saliva was excellent (R=0.99 p<0.001). There was little consistency between the time of sampling and the dose of phenytoin.
In adults, phenytoin has a well-documented S/P ratio between 0.10-0.12 (23) (81) (83) (84) (85) (86), which is in agreement with some of these studies (81) (82). All studies showed high correlations meaning TDM using saliva is an accurate method for phenytoin, however as there is a wide interpatient metabolism associated with it, lack of correlation between drug dose and concentration needs taken into account. The use of stimulated/unstimulated saliva seems to have little effect on correlation and researchers should use EMIT with caution.
Phenobarbital is and older generation AED licensed for many epilepsy types in both adults and children. The pKa of phenobarbital is similar to the pH of saliva and it has been suggested that pH be measured simultaneously (23) however others have questioned this (87). Salivary levels have been shown to be sensitive to changes in pH, which is determined by flow rate (40) (23).
Cook et al (1975) (81) conducted a study on 38 patients (8 months-62 years). Unstimulated saliva (10µl) was collected by spiting/capillary-pipette-dilution system. Blood was collected simultaneously and was centrifuged. No storage conditions were noted. Analysis was conducted by radioimmunoassay and the correlation coefficient was excellent (R=0.98 p<0.001). The slope of 0.857 shows a non-linear relationship between the plasma and saliva. The S/P Ratio was 0.3 and was possibly influenced by pH and drug pKa on the plasma/saliva transfer process. They determined that as plasma levels of phenobarbitone increase, proportionately more appears in saliva. Saliva could be used to determine plasma levels, even though the process was not linear, by multiplying saliva levels by 3.4. The plasma levels were within clinical range and decisions would have been the same for all patients based on their actual plasma levels or the saliva.
In a later study by Mucklow et al (1981) (41), 19 children (<24 months) were studied at 3 weeks, 6 weeks and at 3 month intervals for 1 year. Stimulated saliva was obtained by aspiration under the tongue and blood was collected on the final visit to the clinic. No drug had been taken within four hours of the sampling. The samples were analysed by GC. A highly significant correlation was demonstrated (although r values were not recorded), with a mean S/P ratio=2.2. They were unable to identify a linear relationship between dosage and saliva concentration for phenobarbitone. This was unexpected and probably reflects the fact that, during the treatment period, dosage was adjusted pari passu with body weight and there was therefore little change in weight-corrected daily dosage. Another contributing factor may have been the ionisation>50% of phenobarbitone at plasma pH. Salivary phenobarbitone concentration is sensitive to changes in salivary pH as mentioned before.
A further study by Goldsmith and Ouvrier (1981) (82) collected unstimulated saliva (1-2ml) and plasma from 202 children (5 months-18 years). They used a mucous trap suction device for saliva collection and no decontamination was described. Both matrixes were centrifuged and the supernatant collected for analysis and frozen. They studied two different assays methods, GLC and EMIT. They found that there was better correlation with fresh samples compared to frozen ones. A good correlation was found in the 15 infants (<24 months) with a S/P ratio=0.32 +/- 0.06 and no significant difference was seen between the age groups. For the GLC assay the correlation was very good (R=0.94) and the EMIT was similar (R=0.92). Phenobarbitone is a much more soluble compound than the other AEDs and would thus dissolve in saliva and be removed from the mouth more quickly. The S/P ratio described by this study is similar to that found in adults which varied between 0.30-0.37 (81) (83) (84) (85) (86) no difference was found between age groups but it has been shown that free phenobarbitone levels in neonates are higher than in older children (88) but this applies to new-borns and youngest in this study was 5 months. The study did not consider the time after oral ingestion of the drug and the risk of contamination which was a flaw to the study as they noted that even after 4 hours of chewing their medication, some children had high levels in saliva compared to plasma, possibly due to contamination.
Another study by Gorodischer et al (89) (1997) found poor correlation (R=0.65) which is in contrast to other studies mentioned. This was unexplained after looking at extraction technique of the saliva and potential drug interactions. Phenobarbital is a weak acid with a pKa 7.2 in which the degree of ionisation is sensitive to pH changes near the values of the physiologic plasma pH. As generally only the unionised fraction of a drug crosses biological membranes, in theory phenobarbital saliva concentrations should be corrected for salivary pH (23) (90) but the good correlations were reported without correcting salivary pH (41) (50) (81) (82) (87) (91) (92) (93). Citric acid stimulation may have produced pH changes not only in the saliva but also within the salivary gland, resulting in variable rates of phenobarbital secretion or reabsorption at the level of the acinar cells or the ductal system. However, a study which included stimulation in adults and older children reported excellent correlation (R=0.95) (50). The patients in Knott’s study were older and this was the only major methodological difference, meaning that salivary excretion of phenobarbital may be regulated differently in older and younger children which is in line with this study when they looked at ages of the patients. Those >8 years had the best correlation (R=0.93 plasma total and 0.93 Plasma free). Therefore, salivary TDM of phenobarbital is an accurate method in infants, however the age of the patient must be considered, as too should the flow rate and hence the stimulation which may affect the pH.
Digoxin is a cardiac glycoside which makes the heart beat stronger and with a more regular rhythm. It has been used for more than 230 years and plays a vital role in cardiac healthcare. It is an acidic drug with a pKa of 7.15, thus ensuring that % ionisation will vary significantly with small changes in physiological pH. An increase in pH from 6 to 7 will see the % of digoxin ionised increase from 6.6 to 41.45%.
In infants a fair correlation is seen by Krivoy et al (1981) (94) who compared plasma vs saliva using the RIA digoxin kit. They collected serum and unstimulated saliva (unknown quantity) samples using a syringe from 12 infants six hours after previous dosing with oral digoxin. A S/P ratio of 0.66 +/- 0.20 and a fair correlation co-efficient (R=0.71 p<0.001) was noted, however, it showed a wide range in the ratios. No explanation was provided for the unmatched storage arrangements used during the study. They stated that saliva was an accurate method of TDM, however the poor correlation coefficient does not reflect this.
Furthermore, Berkovitch et al (1998) (25) found good correlation (R=0.87) in 11 stimulated infants receiving oral digoxin. Plasma and saliva (2ml) was collected using a syringe at trough level, and the infant’s mouths were rinsed by water through a bottle. Storage methods were not discussed. Analysis was conducted using fluorescence polarisation immunoassay. They stated that residual digoxin in the oral cavity affected the results, even though the mouths were rinsed. They determined the S/P ratio to be 1.58 +/-1.32nmol/l. They also thought it was possible that digoxin did not enter saliva by diffusion alone, but also by active secretion and that there was a possibility of endogenous digoxin-like substances present in the saliva (95).
Again, a study conducted by Zalzstein et al (2003) (96) showed a good correlation between saliva and serum in 9 stimulated infants receiving oral digoxin (combined with older children n=18), however, the researchers in this study did not agree that salivary TDM of digoxin was an accurate method of TDM compared to plasma. Samples were obtained simultaneously 10-12 hours after dosing. If possible, the children rinsed their mouth with water before the saliva collection. Saliva (1-2ml) was collected from the oral cavity using a mucus extractor. Both sample types were refrigerated at 4oC for up to 24 hours before analysis. Saliva was then centrifuged and the supernatant analysed using competitive fluorescence polarisation immunoassay. A good correlation (R=0.83 p<0.01) was found within a plasma range of 0.1 to 1.1ng/ml. The mean S/P ratio was 2.8, however variability existed in individual S/P concentration ratios. Low plasma concentrations correlated with negligible or undetectable saliva concentrations. No plasma levels were above the agreed toxic range of 2ng/ml, however one saliva was deemed toxic, without the plasma following suit. In three saliva samples the digoxin concentration was not detectable even though the plasma levels were within therapeutic range.
A clear pattern can be seen looking at these studies, the stimulation of saliva yields better correlation, yet further studies will need conducted to determine if, like in adults, TDM using saliva is an accurate method in infants when no stimulation is used. There is also a need for a standardised method of sample collection, storage and analysis before a ratio can be accurately determined as high variability is demonstrated in these studies.
Basic compounds (pKa ≥ 4)
Three basic compounds have previously been studied for saliva content in infants (table 2) with the pKa ranging from gentamycin (10.18) to lamotrigine (5.87). The individual compounds are discussed below.
Gentamycin is a commonly used aminoglycoside antibiotic which is given intravenously (IV) once or twice daily (97). Due to it having a narrow therapeutic index and wide inter/ intrapatient pharmacokinetic variabilities, TDM is necessary to reduce the risk of toxic levels and adverse effects such as ototoxicity and nephrotoxicity (98). It is a hydrophilic drug and with a pKa of 10.18 will be positively charged at physiological pH (99), hence, although it is only weakly protein bound, it would be expected that it does not appear in saliva as Mahmod et al (1983) (100) stated.
In a study by Berkovitch et al (2000) (99), blood and stimulated saliva (2ml) were collected simultaneously before administration of the next dose, using a syringe, no rinsing was necessary due to IV dosing. Storage methods were not described in this study; however, samples were analysed using a fluorescence polarisation immunoassay. A good correlation was noticed in once daily dosing patients (R=0.89 p<0.0001) but no ratio was determined. There was no correlation between TID dosing, this was thought to be due to the physiology of gentamycin transfer into saliva: it was thought that it must take longer for the ionised molecules to enter and equilibrate with saliva and hence OD dosing allows such a time lag as the measurements are taken 24 hours after doing. This study suggests it is a reliable and accurate methods to monitor trough levels in infants on a OD dosing regimen, however the lack of studies on this drug leave the results unreliable and more studies need conducted before an accurate representation of the S/P ratio can be determined
Morphine is an opioid analgesic which exerts its effect on the central nervous system and organs comprising of smooth muscle (38). In infants it is used for terminal illnesses and to relieve respiratory distress associated with end-stage cardiopulmonary disease. It is 30-40% protein bound (101). It is an amphoteric compound, containing both a weakly acidic phenol group (pKa 10.26) and a basic tertiary amine (pKa 9.12). At physiological pH > 98% of morphine will be positively charged.
Kopecky et al (1997) (38) studied 15 children (some of whom were <24 months) receiving IV infusions of morphine. Stimulated saliva (1-2ml) was collected using a mucous extractor in the infants simultaneously after blood collection and frozen at -20oC or lower until analysis using a solid phase serum morphine radioimmunoassay. As the morphine was given IV there was no risk of contamination of the sample by it, however if patients were taking other oral medication their mouths were rinsed with water before sampling. The S/P ratio was 2.28 +/- 2.84. There was no correlation between saliva and plasma (R=0.04 p=0.89). Because citric acid decreased the salivary pH the authors examined the analytical effect of different pH values on the morphine concentrations in saliva across a pH range of 3.96-8.06 and there was no difference in assay performance. In conclusion, determination of morphine levels from saliva cannot be used as a quantitative tool to predict serum concentrations.
Lamotrigine is a newer AED used for a number of seizure types. It acts by blocking voltage dependent sodium channels in the neuronal membrane, therefore reducing the release of excitatory neurotransmitters (22). The metabolism of lamotrigine is influenced by valproic acid, which reduces its elimination through competitive inhibition of the hepatic glucuronidation enzymes, leading to increased serum concentrations of lamotrigine (102) although other drugs can enhance its elimination. It is used unlicensed in infants, with younger children clearing the drug quicker than adults (103). Lamotrigine is a weakly basic compound with a pKa of 5.87.
Tsiropoulos et al (2000) (22) conducted a study on 40 patients >14 years old receiving lamotrigine alone or in combination with other AEDs. They found a linear relationship between the concentration in serum and saliva, stimulated (R=0.94 p<0.05) and unstimulated (R=0.85). And also a strong correlation between the lamotrigine concentration in stimulated saliva and the free lamotrigine concentration in serum after ultrafiltration (R=0.92) or after equilibrium dialysis (R=0.90). In contrast to Mucklow’s suggestion that basic drugs are not affected by changes in salivary pH, this study found there to be a significant difference between the stimulated and unstimulated concentrations, therefore indicating dependence on salivary flow. The excellent correlation they found between saliva and plasma lamotrigine concentrations was better with stimulated saliva and is in accordance with results by (104). They concluded that salivary sampling may be a useful alternative to plasma when monitoring LTG treatment in patients with epilepsy.
A further study by Ryan et al (105) which studied 31 patients (one patient was 2 years old) who had not received their medication within the last 3 hours, nor had eaten/ drank within 15 minutes. They collected unstimulated saliva (0.25ml) by expectorating into a container/ suctioned and blood was collected immediately afterwards. Samples were frozen at -20oC until analysis by HPLC/HPLC-UV. Regression analysis was conducted and best fit equation calculated, with outliers removed. Comparisons were studied using the t-test, Mann-Whitney U test and Wilcoxon signed rank test (p<0.05). The limit of detection was 0.05ug/ml and the limit of quantification was 0.15ug/ml. The correlations were good between saliva and serum concentrations, being tested over two separate laboratories they had R=0.81 and R=0.86. The S/P ratio was 0.62. They found the correlation to be highest in children (R=0.94) and that the ratio was less (0.56). Although they found a wide range between the saliva and plasma concentrations, they agree saliva samples could be used for TDM after ensuring no oral contamination occurs and also noted that caution must be exercised in the collection and analytical aspects of sampling as errors can occur easily.
In a study conducted by Malone et al, 2005 (19), stimulated and unstimulated saliva concentrations were used to monitor lamotrigine therapy in an adult group and in 20 children (1-16yrs) in comparison with serum. The collection of stimulated vs unstimulated saliva was stated as randomised; however, the randomisation method was not described in detail. Saliva was collected twice over thirty minutes and venous blood was collected in-between the two saliva samples. The method of collection nor the storage methods were described in this study, however the stimulation method was using a piece of polymer to chew on, which would not be possible in an infant. Stimulating the saliva in the children only had a small increase on saliva rate from 0.68mL/min to 0.70ml/L. Analysis was conducted using HPLC and statistical analysis was conducted using linear regression analysis and confidence interval determinations. The data in adults showed that within the first two hours following oral administration, the results had a wide scatter, probably due to residue in the patient’s mouths. After exclusion of the results before 2 hours after administration, the results had a high correlation r=0.9841 p<0.0001 (n=98). The saliva to plasma ratio was lower in the adults than children (41.7 +/- 7.07% for unstimulated saliva vs 42.1 +/- 6.52% for stimulated). In children the ratio was 47.6 +/- 7.2% and 46.7 +/- 6.2%. There was a close correlation between the concentrations of lamotrigine in stimulated and unstimulated saliva. The mean S/P ratio=0.49. As plasma lamotrigine concentrations increased from 1mg/L to 10mg/L, the mean saliva to plasma ratio changed from 41.8% to 48.8%, which was statistically significant and shows that saliva to plasma lamotrigine ratios are concentration-dependent. They believe this may be due to plasma protein binding saturation as the drug’s plasma concentration rises. Overall the authors concluded that with appropriate timing, salivary TDM of lamotrigine could provide an appropriate alternative to plasma.
Six neutral compounds have been included within this review. They have been classified as neutral if they have no ionisable functionality or if their pka value is such that a unit change in physiological pH would have less than a 0.1 variation in % ionisation, Table 2. This is further classified as a basic pka less than 4 and an acidic pka greater than 10. The individual compounds are discussed below.
Busulfan is a bifunctional alkylating agent which is used in haematopoietic stem cell transplantation (106). Studies have suggested that a steady state concentration of this drug is vital for successful engraftment in children (107) (108) while high dose is associated with serious organ toxicity (109). It is 32% protein bound (110).
Rauh et al (2006) (55) have demonstrated that TDM of busulfan in saliva is a valid alternative to plasma sampling in infants. Their study involved 10 children including one infant at 13 months receiving oral therapy every six hours. They collected non-stimulated saliva at 0 – 360 minutes after drug administration using a modified medical pacifier which had a perforated teat and filter paper inserted. This method however, would not be suitable for all infants as some do not actually take a pacifier. Samples of 100µg were frozen immediately at -25oC until analysis, however they did demonstrate that busulfan in saliva was stable for up to 48 hours at 4oC (concentration decrease <5%). Analysis was conducted using LC-MS/MS and lower limits of detection and quantification were determined as 2µg/L and 10µg/L respectively. Interference from co-administered antiepileptic drugs was taken into account. The statistical difference between pharmacokinetic parameters in plasma and saliva was analysed by a paired t-test. The method was determined to be linear from 10-2500µg/L (R=0.999 p<0.05). Although the saliva was collected at different time points, no reference was made to contamination of samples following oral administration. The concentration in saliva corresponded to plasma with an excellent linear regression (R=0.958 n=69). They also stated that 15 samples could be analysed in 2 hours with only 30 minutes of labour. The saliva to plasma ratio was excellent at 1.09 and may be explained by the different protein contents or active secretion of the drug into saliva.
The excellent linear regression makes this drug look idea for TDM using salivary samples in this population, however only one infant was involved in the study and therefore further work is required to support the findings.
Fluconazole is a first-line antifungal agent for the treatment and prophylaxis of invasive candidiasis in children and infants (111). It has excellent penetration in tissue and bodily fluids, with 11-12% protein binding (112) and is largely unionized under physiological conditions, which makes it penetrate oral fluid easily. Routine TDM is not normally conducted due to its good safety profile, however infants and children are at a higher risk of suboptimal drug exposure (6).
One study has been conducted on fluconazole TDM using saliva in paediatric patients (56). They studied 10 children 0-24 months taking oral or IV fluconazole, using a suction device to get the unstimulated saliva compared with serum. However, results were collated with all under 18’s totalling 19 patients. Without considering contamination by oral medication the samples (amounts not specified) were collected at trough levels before administration of fluconazole and pair wise after steady state. Blood was centrifuged and serum collected. Both were stored at -20oC until analysis. They showed serum to be stable for 7 days at room temperature while saliva was stable for 17 days at room temperature giving more evidence to the potential for home monitoring. LC-MS-MS was used for analysis and Pearson’s correlation co-efficient was used to determine the correlation between the concentration in serum vs oral fluid, with r=0.96 p<0.1 ratio 1.0. There was excellent linearity across the fluconazole concentration range. Cross validations and bias assessments were conducted. The lower limit of quantification was 0.5ug/ml for both saliva and serum. The oral fluid-to-serum drug concentration ratio did not significantly differ in patients receiving oral treatment versus intravenous treatment (P = 0.791).
In high concentrations of fluconazole, the amount in saliva was less than in serum which suggests saturation of the amount of fluconazole in oral fluid. It has already been established that fluconazole in high doses is well tolerated in paediatrics (113), and in this study only trough levels were considered, therefore saliva sampling could be used to measure trough levels and prevent underexposure. This would also limit the chance of contamination after oral administration. They suggested that salivary sampling in premature infants was difficult and is not suitable due to the lack of saliva production, in contrast it has been suggested to be an accurate method if stimulation is used 5-10minutes before sampling in premature infants (51).
This study shows that salivary TDM of fluconazole in infants is a reliable method to ensure patients do not receive sub-therapeutic levels, which is important in ensuring adequate treatment. Although the population size was small, the correlation was very high and is reassuring that the ratio of 1 is applicable to this population.
Caffeine is frequently used to treat apnoea of prematurity due to lower toxicity levels and a longer half-life than theophylline, meaning longer dosing intervals (114). In a similar way to theophylline, it acts by increasing the central respiratory drive and lessening the threshold of response to hypercarbia (115). It is a drug which also has a narrow therapeutic window and has large patient interindividual variability (116), with metabolism changing with age. There is an increased rate of caffeine metabolism after 36 weeks of age (117) therefore it requires weekly plasma monitoring (51). Caffeine has low protein binding (25-36%) (118) and only the free fraction of caffeine can diffuse across membranes and the total saliva caffeine concentration should therefore reflect the unbound plasma concentration. Salivary glands have a high perfusion of blood, and as caffeine is a highly lipophilic molecule, it can diffuse rapidly into saliva independently of the flow rate and pH of saliva (40) (119), which in theory means the difference between blood and saliva levels should be minor (120).
Khanna et al (1980) (24)conducted a small study on 7 premature infants who received caffeine via an orogastric tube. Blood was collected (150µL) via a heel prick and unstimulated saliva (50 µL) was collected 2-3 times and 4-6 hours after drug administration without considering contamination risk due to oral dosing. Saliva was collected using a sterile rubber ear syringe using suction. They suggested this to be the most convenient method, however failed to advise which other methods were tried. This collection method would no longer be considered appropriate due to swabs specifically designed for the purpose. Both matrixes were frozen until analysis, however the conditions were not specified. Analysis was conducted using straight phase HPLC and regression analysis was conducted with r=0.84 +/-2.8 µg/ml. When the salivary concentrations were <8µg/ml, the serum concentrations did not exceed the therapeutic range of 13 µg/ml and no limit of quantification was determined. The S/P ratio was determined at 1.40. In this study, total methylxantine concentration was also measured as caffeine can convert to theophylline and vice versa (121). In this case r=0.81 +/-3.3 µg/ml. With the exception of one patient, when the total concentration in saliva was less than 8 µg/ml, the total in serum did not exceed 15 µg/ml. The relationship of serum to saliva indicates significant variation when salivary concentrations exceed 8 µg/ml, but below this the serum did not exceed therapeutic range. Although this sounds reasonable, this limit diminishes the usefulness of saliva in this instance as 8mg/L is at the lower end of certain recommended therapeutic ranges for caffeine (122).
In contrast to this Lee et al (123) (1996) conducted a larger study, administering IV caffeine over 7 days to 59 premature newborns (<32 weeks). The infants were randomly allocated (randomisation method not stated) to one of three groups receiving different doses of caffeine. There were 131 pairs of unstimulated saliva (20µL) compared with blood (25µL) which were collected immediately before caffeine administration. Blood was collected via a heel prick or an umbilical catheter and serum was centrifuged. Saliva was collected using another impractical method in newborns- vacuum aspiration from the floor of the mouth. They found their method too labour intensive, taking as long as 30 minutes. Both matrices were stored at -75oC before analysis by HPLC. The limit of quantification was 0.2mg/L (95% confidence intervals). This study did not use the usual correlation coefficient to monitor the strength of their relationships, suggesting that it is a weak statistic in this situation as it simply estimates the degree of association along the regression line and not the line of identity which is of more interest in predictive relationships (124). Instead they used what they felt is a more appropriate and valid measurement of predictive performance which involved the assessment of precision and bias between two methods. The smaller the magnitude of the differences between pairs of measurements, the higher the precision, while a systematic deviation between pairs of measurements can be used to estimate bias. In their study the 95% confidence limits of the RMSE embraced zero, indicating that there was no significant difference in precision between the serum and salivary data.
- 131 samples with a range of 0.28-93.3. precision 0.46 (-0.70-1.12) bias=-2.35 (-2.42- -2.28)
- 84 samples between 8.0-93.3 with a precision of 0.72 (-1.28-2.30) with a bias of -3.53 (-3.66-3.40)
- 41 samples were between 20.0-50.0 with a precision of 0.76 (-1.65-2.79) and a bias of -2.39 (-2.59- -2.19)
In conclusion they found that saliva can be used instead of serum to monitor caffeine at practically any concentration of caffeine even in high maintenance doses of 30mg/kg/day. Their data also suggests that the transport of caffeine from blood to saliva is independent of concentration, meaning it is likely that free caffeine enters salivary ducts by passive diffusion and not via capacity-limited active transport processes that other drugs may use. On the other hand, mean serum caffeine concentrations (29.9mg/L) exceeded the corresponding mean salivary concentrations (27.7mg/L). It was possible that the small amount of bias reflected the fact that it was the total unbound and bound concentrations which were measured in serum whereas only the unbound were measured in saliva. This was unable to be proven however. They also suggested that in breastfed infants, a mother’s caffeine intake may inadequately alter the levels of caffeine in the neonates, therefore rinsing the babies mouth should be considered.
In another large study by De Wildt et al (2001) (51), 140 demographically similar premature infants (<34 weeks) were given an IV loading dose and then continued on oral or IV dosing of caffeine. Blood was collected (200uL) via heel prick or arterial lines and saliva (at least 50µl) collected within 1 hour of routine blood sampling and at least 6 hours after the last caffeine dose. The saliva was stimulated immediately before collection, 5-10 minutes before collection and unstimulated saliva was also collected. The saliva collection method in this study was deemed better compared with the Lee et al (123) method of vacuum extraction, which was time consuming and yielded poor quantities. However, it is far from ideal- gauze attached to a cotton bud, which was placed in the cheek pouch for 5-15 minutes and then syringed- cotton swabs specifically designed for the purpose are now readily available and would offer a much simpler collection method. The author did not mention the contamination risk from oral sampling, nor did they discuss the difference between results due to oral or IV dosing. Blood was centrifuged and plasma separated. Plasma and saliva were analysed immediately or kept refrigerated at 4oC for a maximum of 24 hours. Caffeine concentrations in plasma, saliva and ultra-filtrate were determined using HPLC with UV detection. The analysis method was well described and the limit of quantification was 0.2mg/L with 100µL sample volume which is the same as that described by Lee et al (123). Statistical analysis was conducted using regression analysis.
The unstimulated and stimulated immediately before the collection groups were not well correlated, and it was noted that in more than half of the samples less than 50µl of saliva was collected which may have been a contributing factor to the weak correlation and wide variability. The citric acid stimulated 5-10 minutes before collection had the strongest correlation (r=0.89) and only 8% of the sampling points varied more than 25% from the regression line and the variation in repeatability was small (mean -0.2mg/L +/_ 2SD 2.1mg/L). This means more than 95% of the samples varied less than 2.1mg/L from a duplicate taken within 15 minutes of each other for the whole concentration range. This is adequate for therapy monitoring given the caffeine therapeutic window. Also, applying the citric acid 5 minutes before collection prevented dilution of the sample by citric acid. The authors also ultra-filtrated the plasma for the unbound caffeine levels and found that the correlation between saliva and unbound plasma caffeine concentrations (r=0.76) was no better than the correlation between saliva and total plasma caffeine (r=0.79 immediately stimulated). They determined the mean protein binding to be 30%, which was in agreement with the mean saliva/plasma ratio of approximately 0.7. However, for individual pairs, poor correlation was found between saliva/total plasma ratio and plasma protein binding of caffeine. In conclusion they determined that if the therapeutic window is 10-20mg/L in plasma, then it would be 6-16mg/L in saliva and that this method was reproducible and feasible when saliva was stimulated 5-10 minutes before collection.