*Correspondence: Dr T. Hummel, Department of Otorhinolaryngology, University of Dresden Medical School, Fetscherstr. 74, 01307 Dresden, Germany.
Objectives To investigate whether nausea and vomiting and olfactory sensitivity are correlated, we determined whether subjects with little or no nausea and vomiting are less sensitive to odours than subjects who indicate a high degree of nausea and vomiting, and whether subjects with relatively low olfactory sensitivity are less prone to nausea and vomiting than subjects with relatively higher olfactory sensitivity.
Design Cross sectional study.
Setting The Unit of Perinatal Physiology, Department of Obstetrics, University Hospital, Zurich, Switzerland.
Population Fifty-three women in early pregnancy.
Methods Following a detailed history related to olfaction and nausea and vomiting, subjects filled in a nausea profile which provided a ‘general nausea score’ comprised of the factors ‘somatic distress’, ‘gastrointestinal distress’, and ‘emotional distress’. Olfactory function was assessed using pen-like odour dispensing devices (‘sniffin’ sticks’). Tests included n-butanol odour threshold, odour discrimination and odour identification.
Main outcome measures Olfactory function assessed by means of the sniffing sticks nausea profile.
Results Correlational analyses between results of olfactory sensitivity and scores from the nausea questionnaire were not significant. Further, when subjects were divided into groups with relatively low or relatively high overall scores in the nausea profile, olfactory sensitivity did not differ between groups. Similarly, other analyses did not indicate a modulation of nausea and vomiting through olfactory sensitivity.
Conclusions These findings do not support the hypothesis that higher olfactory sensitivity relates to an increase of nausea. However, they do support the idea that olfactory-induced nausea is independent of subjectively perceived intensity. Olfactory-induced nausea appears to be due to the cognitive processing of olfactory information which, in early pregnancy, is reported to be altered in an unsystematic fashion.
Odorants evoke pleasant or unpleasant emotions. This subjective response differs according to both quality and intensity of the odour. For example, cheese may produce both extremely pleasant and extremely unpleasant responses in different observers. Further, pleasantness may increase with the intensity of some odours (e.g. peppermint) while it may decrease with the intensity of others (e.g. ethyl alcohol)1. Knowledge of the relation between these characteristics is far from complete. Most researchers agree upon the hypothesis that the hedonic response to odorants is not inborn but a product of socialisation2,3. An example of this variability is intercultural differences (e.g. many of the fine cheeses praised by the European connoiseur smell awful to most Japanese4). The anatomical basis of this interaction between olfaction and emotions is the intimate involvement of the amygdala in the processing of olfactory information5,6. This explains, at least partly, why it is possible that odour-evoked emotions are stable and highly reproducible5. Considering the strong hedonic component of most odorous responses, it seems likely that odours may also elicit nausea or vomiting.
In pregnancy, a state of ‘hyperolfaction’ (heightened olfactory sensitivity) is thought to play a role in nausea and vomiting7, which is typically found during the first trimester8. For example, there is an anecdotal report of a woman with congenital anosmia who was not plagued by nausea and vomiting during two pregnancies (B. Prechtl, personal communication). However, substantial research in this area is lacking. Thus, using early pregnancy as a model, the primary aim of the present, prospective study was to investigate whether nausea and vomiting and olfactory sensitivity are correlated. In addition, the following two questions were asked: (A) are subjects with little or no nausea and vomiting less sensitive to odours than subjects who indicate a high degree of nausea and vomiting, and (B) are subjects with relatively low olfactory sensitivity less prone to nausea and vomiting than subjects with relatively higher olfactory sensitivity?
For this study, 53 pregnant women [age 28 (5.6) years, mean (SD)] were recruited after the protocol had been explained in great detail. All subjects gave written consent. The study was performed according to the Declaration of Helsinki regarding biomedical research involving human subjects. It was approved by the Ethics Commitee of the University of Zurich Medical School. Pregnant subjects were tested in early pregnancy before 14 weeks of gestational age (see Table 1). All subjects were in good health. They were instructed to refrain from smoking, drinking coffee and eating at least 1 hour before the study. Exclusion criteria were severe diseases (e.g. endocrinologic diseases and epilepsy), drug and nicotine abuse (more than 10 cigarettes per day) and known olfactory or gustatory dysfunction.
Table 1. Descriptive statistics across all 53 participants.
Min = minimum; Max = maximum; SEM = standard error of mean.
Week of pregnancy
Overall nausea score
Butanol odour threshold
Prior to examination of olfactory sensitivity, subjects were asked to fill in a questionnaire with questions related to general olfactory abilities and, specifically, nausea and vomiting during the last three days. Moreover, subjects filled in a nausea profile9. This nausea profile consisted of 17 questions which provided a ‘general nausea score’; in addition, it allowed the discrimination between factors contributing to the general nausea score, namely, ‘somatic distress’, ‘gastrointestinal distress’ and ‘emotional distress’.
For assessment of olfactory function, pen-like odour dispensing devices (‘Sniffin’ Sticks’) were employed10,11. This kit comprises three tests of olfactory function, namely, tests for odour identification, odour discrimination and n-butanol odour threshold. For odour identification, 16 odorants were presented to each subject who were free to sample the odours as often as necessary in order to identify the odours from a list of four descriptors. The experimenter presented odour pens separated by an interval of at least 30 seconds to prevent olfactory desensitisation. Odor discrimination was performed using 16 triplets of odorants. Subjects were presented with three odorants; their task was to identify the sample that smelled differently. Subjects were blindfolded to prevent visual detection of the target odour pens. They were allowed to sample each odour only once. Presentation of odour triplets was separated by at least 30 seconds. The interval between presentation of individual odour pens was approximately 3 seconds. Odor threshold was determined using n-butanol as the odorant; dilutions were presented in a geometric series. Presentation of the odorants was similar to that described above for the discrimination task. Again, subjects were blindfolded to prevent visual identification of the odour-containing pens. Three pens were presented to each subject in a randomised order, one contained the odorant at a certain dilution, the other two solvent only. The subject's task was to find out which of the three pens smelled differently. Presentation of triplets occurred every 20 seconds, until subjects had correctly discerned the odorant in two successive trials which triggered a reversal of the staircase. From a total of seven reversals, the mean of the last four staircase reversal points was used as threshold estimate12.
Data were analysed using SPSS 10.0 for Windows. Group results were compared using the Mann–Whitney test. In addition to correlational analyses using Spearman statistics, differences between the following groups were investigated: (Analysis A) Subjects with relatively low or relatively high scores in the overall nausea score; (Analysis B) Subjects with and without presence of nausea and vomiting during the last three days according to the subjects' answers in the questionnaire; and (Analysis C) Subjects with relatively low or relatively high scores in the three olfactory tests. For Analyses B and C, the medians of the respective parameters (overall nausea score, results of olfactory tests) were used to divide subjects in two groups. The alpha level was 0.05.
Descriptive statistics across all participants are presented in Table 1.
Correlational analyses between results of olfactory sensitivity and scores from the nausea questionnaire did not produce significant results (r53 < 0.21, P > 0.14; Fig. 1). In addition, when subjects were divided in two groups with relatively low or relatively high overall scores in the nausea profile (Analysis A), olfactory sensitivity did not differ between groups (Z < 0.53, P > 0.59). Also, Analysis B revealed that subjects with little or no nausea and vomiting three days before olfactory testing were not significantly different in terms of olfactory sensitivity than subjects who indicated the presence of nausea and vomiting (Z < 1.51, P > 0.12). Finally, Analysis C indicated that general nausea scores did not differ between subjects with relatively lower or higher scores in the three individual olfactory tests (Z < 1.65, P > 0.11).
The present data indicate that olfactory sensitivity has little, if any, impact on nausea and vomiting during early pregnancy. This result is not what one would expect from the literature. For example, Erick7 talks about ‘hyperolfaction gravidarum’, which might be related to ‘hyperemesis gravidarum’ (pernicious vomiting of pregnancy)—although the experimental basis for such a statement is very small. The above notion is largely inferred from (a) evidence/hypotheses that olfactory perception is correlated with hormonal changes during the menstrual cycle13,14, and increased vascularity in the respiratory tract as a consequence of these hormonal changes15, and (b) anecdotal reports of a drop of olfactory sensitivity in postmenopausal women without oestrogen replacement16.
However, systematic experimental studies investigating changes of olfactory function during pregnancy have revealed little evidence for an increase in olfactory sensitivity during pregnancy17–19—while changes in hedonic ratings of odours are present, but do not present themselves as a generalised phenomenon related to all odours18–21. In other words, olfactory changes during pregnancy appear to relate mostly to changes of the cognitive processing of odorous information, but not to changes in olfactory acuity.
The present results do not exclude the facts that (1) the cognitive processing of odorous sensations appears to change during early pregnancy18–21, and that (2) many odours may trigger nausea and vomiting especially when considering the different hedonic perception of odours during pregnancy. For example, odours appear to play an important role in the anticipatory nausea which is frequently observed during chemotherapy22–24. In addition, data from Mattes and Cowart25 indicate that the prevalence of food aversions are particularly low in anosmic patients (approximately 11%) compared with the general population (approximately 26%). They interpreted their findings such that this might reflect the ‘importance of an odour as a signal about a food's postingestive effects’.
Taken together, our findings support both the hypotheses that higher olfactory sensitivity is not related to an increase of nausea, and that olfactory-induced nausea is independent of subjectively perceived odour intensities. Olfactory-induced nausea rather appears to be due to the cognitive processing of olfactory information which, in early pregnancy, is reported to be altered in an unsystematic fashion.