Concord Dependency Seminars were previously published at

22 May 2007

The Interpretation of Urine Toxicology in Dependency Treatment. Principals and Pitfalls.

Concord Seminar 22 May 2007

The Interpretation of Urine Toxicology in Dependency Treatment. Principals and Pitfalls

Speaker: Dr John Lewis, Toxicology Unit, Pacific Laboratory Medicine Services

Dear Colleagues,

Dr Lewis is one of the world’s leading experts in drug testing. His speaking manner combines what T. S. Eliot might have termed a lugubrious drollery with a profound grasp on his subject. It is easy to be light-hearted about ‘piss tests’ but it is also deadly serious if your own job, drivers licence or liberty depend upon such a result.

We were reminded first up what urine testing can NEVER determine with any accuracy: (1) the dose, (2) the time it was taken or (3) the pharmacological effect of any substance being tested.

The most common drug assays they perform are for methadone and metabolites, cannabinoids, opiates, cocaine, benzodiazepines and amphetamines. Barbiturates often omitted these days since their illicit use seems to have ceased for all practical purposes. The term ‘amphetamine type substances’ (ATS) is now superseding ‘sympathomimetic amines’. This group includes dexamphetamine, methylamphetamine, ecstasy (MDMA), methylenedioxyamphetamine (MDA), and other ‘designer’ drugs such as paramethoxyamphetamine (PMA) and their metabolites, but also ephedrine, pseudoephedrine. One needs to know the particular immunoassay ‘kit’ being used to be sure what exactly is detected and at what level.

Laboratories are asked to perform tests both in a clinical setting as well as for forensic, workplace or medico-legal reasons. For clinical purposes a cost effective and fast turn-around time approach is used. This starts with an inexpensive immunoassay which is very sensitive for most of the drugs being tested for, but generally not specific. Hence a negative batch of tests can yield a fast, efficient response to the clinician. Positive immunoassay results for any of the drug groups (or negative for methadone) may indicate further testing, typically using GCMS (gas chromatography/mass spectrometry), which is considered the ‘gold standard’. Although thin layer chromatography (TLC) is not commonly used nowadays, Dr Lewis says it still has a place: it presents information on a large range of drugs to view at a single glance, and is inexpensive. Because the TLC depends upon the human factor of recognising patterns, it is subjective and unless the spot patterns correspond to known medication, confirmatory testing by mass spectrometry is usually conducted. Although it is not used for medico legal work, it still has a place in clinical settings, as an adjunct to mass spectrometry in presumptively identifying a wide range of therapeutic substances not amenable to immunoassay.

In particular cases there will need to be specific tests done, especially for suspected drug use which may not be detected by the usual immunoassays. These include tests for doctors, nurses or other health care workers on conditional registration due to drug use. Such drugs include pethidine, tramadol and the short acting anaesthetic propofol. Abuse of these drugs outside the medical setting is exceptional.

Note that buprenorphine is also hard to detect by simple methods. Although there is an immunoassay for the drug, toxicologists must be aware of possible false positives from a number of unrelated therapeutic substances. However, like methadone, when the dose is taken under supervision such testing is less important than, say, in England where much treatment is unsupervised and testing for the prescribed medication can be crucial in determining compliance and overall stability.

Dr Lewis then detailed the limitations and strengths of modern immunoassays in determining a class of drug but only in two cases can they detect specific metabolites, EDDP (for methadone) and 6-mono acetyl morphine (heroin). The value of a negative test was pointed out. We were reminded that testing was almost pointless in hospital casualty cases: for overdoses, the results are usually not available until either the patient is dead or else recovered. Also, medications are used so routinely and such patients may have injuries necessitating local anaesthetics, dressings, iodine, etc in the course of their treatment in the casualty ward that results are close to meaningless.

Specifically, Dr Lewis said that positive opiate and ATS immunoassays should be taken with caution as there are many causes of false positives. These include poppy seeds, cough mixtures, decongestants and common analgesics. Dr Lewis told us that his own urine remained positive for ‘opiates’ for nine days after a dose of the cough suppressant pholcodine. The main value of these screening tests is when the result is negative. Note that ‘opiate’ immunoassays do not detect the ‘opioids’ methadone, buprenorphine, pethidine and others. Oxycodone has only a very weak response to “opiate” immunoassays.

We were then shown the plates used for thin layer chromatography and a list of 20 common drugs which can be definitively determined using this method (eg. morphine, codeine, oxazepam, pseudo-ephedrine, paracetamol and nicotine). GCMS was then described in response to a question from the floor. In essence it appears that there are two properties of each substance which are identified in the method, causing a unique fingerprint from the two derived figures. It is more expensive than other methods, but more accurate and specific, being able to detect both the original base compound as well as ‘derivatised’ products.

Then we had a brief tutorial on the use of testing for alcohol consumption. Everyone knows about breath testing, but 5% of alcohol is excreted in the urine and there is a direct correlation between plasma and urine alcohol concentration of 1.3:1. However, due to the short half life of alcohol, such testing is only of any use within hours of the drug use. And, as with other drugs, a certain level could be associated with a small amount of drug used very recently, or equally, a large amount used quite some time before.

There are also unexpected false positives, including a case Dr Lewis described where urine from a diabetic in a rehabilitation facility had undergone fermentation (probably by yeasts) before being tested; the calculated blood alcohol concentration (0.34) would have been lethal. A less ‘gross’ error might not have been discovered, and this would have led to the automatic expulsion of the person from the rehab facility.

Tests for cannabis are of limited value since, for most, its use is not relevant to the treatment or supervision being given. Hence Dr Lewis only performs cannabis tests when specifically requested, such as in patients being treated for cannabis dependence, to assess progress.

We were then taken through some metabolic pathways. Heroin breaks down within minutes into 6-acetyl morphine, then to morphine. This then is broken down into morphine-6 or -3 glucuronide which are excreted. Codeine is largely conjugated into codeine-6 glucuronide, but importantly, a small proportion is transformed into morphine. A positive test for morphine can therefore sometimes occur due to codeine use (but not the other way around). A warning: most tests underestimate the amount of codeine in urine, as the metabolite codeine-6 glucuronide is hard to "bust" into codeine, which can be detected. It is important to know the relative amounts of morphine and codeine in a urine sample as the ratios affect correct interpretation as to what may or may not have been ingested.

Diazepam is broken down into another active metabolite, oxazepam. This can occur via two intermediaries, nordiazepam and temazepam. Most of the common sedatives and related drugs such as clobazam will show up as benzos on the initial immunoassay. However, specific confirmatory testing must be done when clobazam is used in therapeutic trials to test against ‘street’ benzos.

Stimulants were then covered including the new definition of ice in an age of global warming (ice-bergs and all!). Amphetamine was first synthesised by the Germans in 1887. It was heavily marketed in the US in the 1930s as ‘Benzedrine’. Methylamphetamine is easier to manufacture, especially if one has the base product pseudoephedrine. We were then told that the latest ‘craze’ for stimulants is purely based on stronger, highly purified drug being available in the form of ‘crystal meth’ or ‘ice’. Methylamphetamine powder is a salt, "crystal" a highly purified salt, and "base" is an oil. Urine testing cannot distinguish between them as these are the same drug. While Dr Lewis’ lab has found 2005 was the year with highest mean amphetamine levels, in 2006 the maximum levels found each week continued to climb to being 5 fold the 2003 levels. While these are dramatic findings, it is hard to know their significance overall except to imply that some users are taking very large amounts of methylamphetamine, viz, “ice”.

Cannabis has many metabolites which are detected on screening, and confirmed with carboxy –THC on GCMS. It is very lipophilic, and gets stored in the fat cells of the body. Cannabinoid urine tests may be negative within a few hours of a single smoke; daily use may take many days, and heavy use a month or more. If a high level is found then it is easy to know that there is continuous use. Carboxy-THC: creatinine ratios can indicate increasing or decreasing use (see case vignettes below).

Then there was a discussion about laboratory ‘cut-offs’ which are essential for legal purposes, but less meaningful for clinical purposes, except to reduce the numbers of false positives. Cut-offs are also necessarily somewhat arbitrary, like the drink driving limits - and can vary from place to place or from time to time. Currently 50ng/ml is used for immunoassays of cannabinoids, and 15 ng/ml for the specific GCMS for carboxy-THC (plus or minus a figure for lab uncertainty; this means an actual cut off of around 18-19 ng/ml). Dr Lewis believes there is a case for higher cut-offs to be used for cannabinoids, to identify substantial cannabis use, rather than low level or more importantly residual drug from previous heavy use.

Some case vignettes in the second half illustrated common problems. Three patients with positive immunoassay for opiates claimed only to have taken codeine-based analgesics. One had codeine and morphine on GCMS, and this could be explained by metabolism of codeine to morphine, or other sources or morphine such as poppy seeds, morphine sulphate etc. Another had urine positive for morphine, and negative for codeine: this could occur if there was extensive metabolism of codeine to morphine (for example by cytochrome CYP2D6 ultrarapid metabolisers) and especially if the laboratory test underestimated the amount of codeine (see above). In the last case, urine was positive for morphine and monoacetyl-morphine: the latter can only come from heroin use.

In a case of roadside drug testing, a woman justified her positive salivary cannabis test by saying "I never smoke pot, but my partner smokes it all the time". Dr Lewis explained that this test does not pick up metabolites of THC, only the parent drug, and is not very sensitive, missing a large proportion of cannabis users (as reported by the European ROSITA study). Thus passive smoking could not cause a positive test. A man on methadone, who had not had a positive urine test for many years, blamed his positive urine cannabinoid test on his partner, who ‘smoked 30 cones each day’. A positive immunoassay test result is unlikely to be a result of passive inhalation. It is more likely be a false positive due to other medication, cross contamination or else laboratory error.

Dr Lewis described the benefits of using carboxy-THC:creatinine levels to help allow for variation in urine concentrations due to level of hydration. Cases were shown from the Drug Court, where declining THC:creatinine ratios were consistent with ongoing abstinence; in another case a spike in ration of THC:creatinine led to punitive action, but might have been explained by the person going to the gym, and mobilising cannabinoids stored in fat cells.

Another case from the Drug Court showed how the sequence of appearance or disappearance of diazepam metabolites (nordiazepam, oxazepam and temazepam) could be used to make inferences about recent diazepam use. In this case, as in almost every example discussed, Dr Lewis was able to give examples of exceptions, where other causes than the most obvious might account for the result. So urine tests should never be interpreted uncritically by untrained people.

In another case, a worker was suspended for producing "dilute urine" (wrongly described as a "false negative urine test") because of low creatinine urine (1.4 mmol/L), and allegedly told he would need to produce two urine tests with creatinine higher than 5 mmol/L. However, this worker's serum creatinine was low owing to lean body build, while urea, electrolytes, specific gravity, osmolality were consistent with physiological urine. THC:creatinine ratios might help adjust for hydration (some people deliberately drink lots of water to dilute their urine) but could also discriminate against people with naturally low creatinine. A urine creatinine level as low as 0.9 mmol/L is physiologically achievable. Below this suggests the likelihood, and below 0.5 mmol/L the near certainty, of external interference with the sample, usually meaning dilution after urination.

Laboratories and clinicians need to be careful with the information they give as employers may misinterpret loose terminology.

Comments by Andrew Byrne, Richard Hallinan and Judith Meldrum, from Dr Lewis’ talk and power point presentation.