By: 3 April 2013


The olfactory sense is yet another means by which diagnosticians can find clues about ailments or uncover little fibs embarrassed patients might be telling. Dr Bridget Osborne warns us not to turn up our noses at using smells in diagnosis.





I n the twenty-first century we are subjected to relentless olfactory stimulation. Simply walking down the High Street, we are enveloped by smells, which we barely register. We pass fragrant coffee bars, delicious-smelling restaurants, bakeries with the wonderful smell of freshly baked bread, and our fellow shoppers, each with their individual scent of aftershave or perfume.

These ‘loud’ smells, many of them artificial, can desensitise us to the underlying odour of human kind, including the smell of earwax, sweaty armpits, feet and menstrual blood. Compared to other animals, such as dogs, our sense of smell is almost vestigial. However, although we have only six million olfactory receptors, compared to 300 million in dogs,1 we are still able to detect substances in dilutions of less than one part in several billion parts of air.

We do not have the vocabulary to describe smell in the same way that we are able to describe colour – for example we all understand that blue can be anything from navy through petrol to the palest duck shell blue. However, to the majority of people, vinegar smells of vinegar, although cooks and connoisseurs are able to differentiate balsamic from malt. Much like a wine taster educates his or her palate to fine wine, distinguishing all the different notes, all clinicians can educate themselves to detect smells in a clinically useful way too.

The primary olfactory cortex is linked to the amygdala and hippocampus, which are involved with emotional and long-term memory, with the result that certain smells can provoke vivid recall of people and events. Olfactory memory may well be the trigger that makes experienced clinicians worry about a patient, even when they cannot clearly identify their concerns. What they are doing is subconsciously remembering a different patient from the past, who had a ‘bad smell’.


A nose for pheromones

In addition to smelling with their noses, animals as diverse as cattle and snakes have an accessory olfactory system, known as the vomeronasal organ, located in the roof of the mouth. This is primarily used to detect pheromones, rather than the inhaled chemical molecules detected by the nose. The presence of a similar structure in humans was doubted for many years, but its existence has recently been histologically proven in about 60 percent of people, using cadaveric dissection.2 Animals using their vomeronasal organ show a ‘Flehman response’, demonstrated, for example, when a ram following a receptive ewe can be seen to lift his head and curl back his top lip.

Pheromones, also known as ecto-hormones, are substances used for olfactory signalling throughout the insect and animal world. The first pheromone to be isolated was bombykol, a chemical produced by the female silkworm moth (Bombyx mori) to attract a mate. There has been much controversy about human ability to either produce or detect pheromones, but research has increasingly shown that not only do humans produce these substances, but that they also influence human behaviour, including sexual and social relationships.

This research has been exploited commercially and a brief browse of the internet will show numerous artificially produced substances available. One example of an online advert for a product we’ll call Product X, reads: “Product X is intended to attract women by helping you radiate pure alpha dominance. It helps create gut level attraction, by emitting the pheromone signature of an aggressive, sexually active young man who gets what he wants. It makes women feel butterflies in their stomach any time they talk to you. Product X gets its best results with women who are naturally attracted to aggressive ‘bad boy’ types, and looking for fun.”

Substances which may act as pheromones are produced by apocrine glands, which are found principally in the axillae, areola and genitalia in humans. Almost 400 different volatile compounds have been identified in axillary sweat using gas chromatograph-mass spectrometry, which have individually distinct and reproducible spectrometric fingerprints with reproducible differences between the sexes.3

Short chain fatty acids, known as ‘copulins’, whose composition varies throughout the menstrual cycle, occur in vaginal secretions. It has been suggested that there are two types of pheromone: ‘signal pheromones’, which produce short term behavioural change, and ‘primer pheromones’, which activate the hypothalamic pituitary axis and trigger release of GnRH, inducing longer-term changes through subsequent release of gonadotrophins.4

Although humans are largely unaware of being able to detect pheromones, we undoubtedly are able to do so, as shown by the synchronisation of menstrual cycles in groups of women living or working in close proximity.5 Work in mice has replicated this effect, but also shown other interesting effects, such as earlier maturation of females in the presence of an unrelated male, known as the ‘Vandenbergh effect’, as well as the rather more sinister ‘Bruce effect’, in which pregnant mice housed with a male that was not their original mate had significantly more miscarriages before mating with the new male.4

There are also a number of studies looking at the effect of androstenol on male sexual attractiveness to females, but the effects seem to depend upon the phase of the menstrual cycle and to be blunted by oral contraception.4

Another interesting effect of hormonal attraction is that there appears to be olfactory signalling related to immuno-competence, with women finding the odour of men from dissimilar HLA types more attractive than that of those with similar HLA types.4 However, this effect is reversed in women taking oral contraception, which may be due to the effect of the pill mimicking pregnancy, as pregnant females are more likely to be attracted to close kin who are ‘safer’ from a reproductive standpoint.


The tells of smells

Antenatal ward

In the antenatal setting, we should all be aware of the social circumstances of our patients. The smell of nicotine ought to elicit advice about smoking and the dangers of exposing a newborn to smoke. Even when the pregnant mother herself denies smoking, she may live in a house with other smokers to whose fumes her baby will be exposed.

Cannabis users have a lingering sweet and dusky smell, while the smell of damp and cabbage evokes poverty and should prompt gentle inquiry into home circumstances, the search for poor nutrition, anaemia and a small-for-its-dates foetus. Peppermint is a signal to enquire about heartburn, or to question what it is perhaps intended to hide, such as nicotine or alcohol.

Ladies with hyperemesis may carry the scent of vomit on their breath in addition to the smell of ketones, which are produced by the liver when its store of glycogen is exhausted. Not everyone has the ability to smell ketones, which are a sweet smell, like very ripe apples or nail polish, which is why dip-testing urine is so important in this situation.

The onset of labour may be announced by an obvious membrane rupture, but an amniotic fluid leak may be confused with urinary incontinence, and many websites intended for pregnant women suggest smelling the fluid to differentiate the two. Some liken the odour of amniotic fluid to that of semen, while others describe a sweet smell, unlike the acidic smell of urine. If there has been an undetected leak and chorioamnionitis has developed, there will be a foul smell of anaerobic bacteria, which should prompt urgent action.

Although thankfully uncommon in women, many sheep farmers will be all too aware of the smell of corruption, when a ewe is carrying a putrefying dead foetus, which sometimes happens when listeriosis occurs due to ingestion of soil-contaminated silage.

All new mothers love the smell of their newborn and will linger over the soft, downy head, imbibing the newborn smell. Although this is the smell of milk and baby bath, there is also a special, individual scent.  We know that sheep recognise their offspring by smell. At feeding time in spring, sheep will race up to the trough, leaving their lambs behind. After feeding, the ewes find their lambs by sniffing them, butting away any which are not their own. Lambs themselves are less fussy and will suckle any ewe which will allow them to.

Because sheep will reject a lamb which is not their own, farmers fostering lambs rub a lamb which is to be adopted with the ewe’s own membranes or placenta, use perfume to disguise its smell, or place the skin of the ewe’s own dead lamb over it, completely covering the rump, until the ewe’s own milk has gone through the lamb’s digestive tract for two or three days.

In humans also the baby’s own smell is vital to bonding and it is noticeable that many mothers will feel compelled to bathe their baby after it has been nursed by a stranger wearing ‘foreign’ perfume. Human babies, unlike lambs, do prefer their mother’s smell, with babies favouring breast pads worn by their own mothers. Babies also seem to have a brief window of olfactory learning after birth when they are able to learn odours, again promoting bonding with their birth mother.6

The smell of a newborn may also be important diagnostically, as several metabolic disorders cause a change in the smell of the skin and urine. Maple syrup urine disease is a neurodegenerative disorder caused by an enzyme defect in the handling of the amino acids leucine, isoleucine and valine, in which the baby’s urine smells of burnt sugar.7

In phenylketonuria, phenylalanine accumulates and is converted to phenylketone, which appears in the urine, and can be smelt on the skin and hair giving a musty or mousy smell,8 whilst isovaleric acidaemia, a disorder of leucine metabolism, causes an odour reminiscent of sweaty feet.9 Tyrosinaemia results in the accumulation of phenylalanine, tyrosine and sometimes methionine, and babies with this disorder are said to smell of boiled cabbage.

All babies in the UK are screened for phenylketonuria with the Guthrie test,10 but this may not detect all disorders of metabolism, such as tyrosinaemia, which is found in some Asians living in the West Midlands.11 It is therefore important to take seriously any mother who says her baby “smells funny”, as early dietary intervention and restriction may prevent future pathology developing.


Gynaecology clinic

A good sense of smell may also be helpful in the gynaecology clinic. Many women diagnose their own recurrent urine infections by the smell, as much as by their other symptoms; unsurprisingly, cultures of E. Coli smell like faeces. Stale urine has more of an ammoniacal smell, which alerts us to urinary incontinence and should provoke sensitive questioning, especially in older ladies.

In general practice, it is not unusual to be called upon to retrieve a retained tampon, an act which may necessitate evacuation of the room for a while afterwards, or generous use of air freshener. We are also taught that vaginal irritation associated with a ‘fishy’ odour is likely to be the result of vaginosis, where the normal Lactobacillus species in the vagina, which produce hydrogen peroxide, are replaced by anaerobic bacteria such as Prevotella sp, Mobiluncus sp, G. Vaginalis, Ureaplasma and Mycoplasma.12 There is said to be a characteristic smell of advanced cervical cancer, which may also be a result of the proliferation of anaerobic bacteria.


Other diagnosable smells

Awareness of smell is useful in other areas of medicine, for example the foetor on the breath of a patient with appendicitis or the characteristic foetor hepaticus, a musty smell caused by the breakdown of excess methionine in liver failure. In renal failure, urea in saliva breaks down to ammonia, a smell which is combined with the fishy smells of dimethylamine and trimethylamine, while blood in the gut causing melena also has a characteristic smell.

At present there are few useful diagnostic tests of smell, although gas chromatography has been used to detect the chemicals in patients with hepatic failure.13 However, turning once more to animals, dogs, using their extraordinary olfactory ability, have been trained not only to detect drugs and bombs, but also to sniff out bladder and lung cancers.14 Recently, dogs have been trained, using perspiration samples from hypoglycaemic patients, to alert carers to detect diabetic hypoglycaemia. This is especially helpful for parents of children with Type I diabetes, where hypoglycaemic episodes may be unpredictable.15



Smell is a difficult topic to teach and often overlooked in training, as it is a subjective sense and difficult to describe, other than by using comparisons. All too often, ‘natural’ smells are disguised by artificial perfumes, air fresheners and talcum powder. However, we should all learn to use our noses at the bedside, in just the same way as we listen to our patients, look at them and touch them, as smells can convey useful clinical information and are an adjunct to our diagnostic armoury.

It is time to put the nose back into diagnosis, and all at no extra cost to our troubled NHS. 



  1. G. Mather. Foundations of perception 2nd edition, Psychology press Chapter 2: The chemical senses, Smell (pp. 39–44)
  2. Won J, Mair EA, Bolger WE, Conran RM  The vomeronasal organ: an objective anatomic analysis of its prevalence.. Ear Nose Throat J. 2000 Aug;79(8):600-5.
  3. Penn et al, (2007). Individual and gender fingerprints in human body odour. J. R. Soc. Interface 4, 331–340
  4. Karl Grammer, Bernhard Fink, Nick Neave,Human pheromones and sexual attraction. European Journal of Obstetrics & Gynecology and Reproductive Biology Vol 118, Issue 2 , Pages 135-142, 1 February 2005
  5. Pheromones regulate ovulation BMJ 1998;316:797 (Published 14 March 1998)
  6. Romantshik et al Preliminary evidence of a sensitive period for olfactory learning by human newborns. Acta Paediatrica.Volume 96, Issue 3, pages 372–376, March 2007
  7. Menkes JH, Hurst PL, Craig JM. A new syndrome: progressive familial infantile cerebral dysfunction associated with an unusual urinary substance. Pediatrics. Nov 1954;14(5):462-7.
  10. Guthrie R, Susi A, Simple phenylalanine method for detecting phenylketonuria in large populations of newborn infants. Pediatrics.1963 Sep:32:338-343
  11. C. Hutchesson, S. K. Hall, M. A. Preece, and A. Green Screening for tyrosinaemia type I ,Arch Dis Child Fetal Neonatal Ed. 1996 May; 74(3): F191–F194.
  13. Van den Velde S, Nevens F, Van Hee P, van Steenberghe D, Quirynen M.J Chromatogr B GC-MS analysis of breath odor compounds in liver patients AnalytTechnol Biomed Life Sci. 2008 Nov 15;875(2):344-8. Epub 2008 Sep 17.
  14. Michael McCulloch, Tadeusz Jezierski, Michael Broffman, Alan Hubbard, Kirk Turner and Teresa Janecki Diagnostic Accuracy of Canine Scent Detection of Early- and Late-Stage Lung and Breast Cancers. Integrative cancer therapies 5(1); 2006 pp. 1-10
  15. Wells DL et al. Canine responses to hypoglycemia in patients with type 1 diabetes. Journal of Alternative and Complementary Medicine 2008 Dec; 14(10): 1235-41.