Introduction

The present study was undertaken to address the concern thatimmunosuppression may occur in Antarctic expeditionersand be associated with the anecdotal observation of anincreased incidence of infections in expeditioners when iso-lation is broken. The effect of cold exposure and isolation onthe immune system of expeditioners who winter in Antarcticahas been the subject of limited research that has focused onsystemic immunity. The very nature of the Antarctic condi-tions has restricted the investigations to studies of humoralimmunity,1,2 delayed-type hypersensitivity using the CMImultitest system2–4 and more recently T-cell proliferationresponses to mitogens.5 The epidemiologies of several viral5,6

and bacterial7 infections and responses to vaccination2 havealso been examined during polar isolation. These studies haveshown that prolonged Antarctic isolation is associated withalterations in T-cell function, resulting in decreased immuno-logical responsiveness,2–5 lowered vaccination antibodyresponses,2 increased viral shedding of latent viruses5 andsevere responses to viral challenge.6

The aim of the present study was to investigate the impactof Antarctic isolation on mucosal immunity. Mucosal

immunity, in association with innate non-specific factors,forms the first line of defence against antigens presented atmucosal surfaces.8 Secretory IgA (SIgA), and to a lesserextent secretory IgM (SIgM), antibodies play the major rolein effective specific immunity.8 Immunoglobulin G plasmacells are in low numbers at mucosal sites, but serum-derivedand locally produced IgG antibodies are important in pro-tection of the respiratory tract.9,10 Alterations in IgG andalbumin concentrations in mucosal secretions providemarkers for increased membrane permeability, dehydrationdue to flow rate changes and sample deterioration.11

Saliva has tended to be the mucosal secretion of choice formost studies of the human mucosal immune system due tothe ease and standardization of collections.11 Salivary IgA isa well-accepted measure of activity of the mucosal immunesystem and has been used to characterize the effects of manystressors. The integrity of mucosal immunity can be com-promised by many factors,11 several of which affect expedi-tioners during an Antarctic winter. The Antarctic environmentof isolation, cold and photoperiodicity is the most extreme on Earth.12 Stress, whether physical, emotional or chemical,triggers a complex series of host responses encompassingalterations in immunity and reduced resistance to infection.13

The stressors in Antarctica are amplified by isolation andanxiety and have been shown to have a negative effect onCMI response.2

The effects on mucosal immune parameters were assessedduring the winter of 1992 at three Australian Antarctic

Immunology and Cell Biology (2000) 78, 616–622

Research Article

One year in Antarctica: Mucosal immunity at three Australianstations

M GLEESON, 1 JL FRANCIS, 2 DJ LUGG, 3 RL CLANCY, 1 , 2 JM AYTON, 3 JA REYNOLDS 3

and CA MCCONNELL 3

1Hunter Immunology Unit, Hunter Area Pathology Service, Royal Newcastle Hospital, Newcastle, 2Faculty of Medicine and Health Sciences, University of Newcastle, Callaghan, New South Wales and 3Polar Medicine Branch, Australian Antarctic Division, Channel Highway, Kingston, Tasmania, Australia

Summary The effect of a year’s isolation in Antarctica on the human mucosal immune system was assessedduring the winter of 1992 at three Australian Antarctic stations: Casey, Davis and Mawson. Saliva samples werecollected from each expeditioner prior to their departure from Australia and during each month in Antarctica. Theconcentrations of salivary immunoglobulins IgA and IgG were significantly different between the three stations,but there were no differences for salivary IgM and albumin. The mean concentrations of IgA were higher atMawson (P < 0.008), and the mean concentrations of IgG were lower at Davis (P < 0.001) compared with the otherstations. Ranges of values observed at the stations over the 12–13 months were similar. The variability of valueswithin individuals showed station differences for salivary IgM and IgG only. The study revealed significant changesin salivary immunoglobulin values over the period in Antarctica, with similar patterns at the three Australian stations. The salivary IgA and IgM levels were lower in the first 4 months in Antarctica (January–April) andincreased to maximum values in July–August, before returning to mean levels when isolation was broken inOctober–November. The patterns of salivary IgA and IgM suggest that stressors due to isolation may play a rolein alterations of mucosal immunity in expeditioners in Antarctica.

Key words: albumin, Antarctica, immunoglobulin A, immunoglobulin G, immunoglobulin M, mucosal immunity,saliva, stress.

Correspondence: Maree Gleeson, Hunter Immunology Unit,Royal Newcastle Hospital, PO Box 664J, Newcastle, NSW 2300,Australia. Email: mgleeson@mail.newcastle.edu.au

Received 23 February 2000; accepted 6 June 2000.

stations: Casey, Davis and Mawson. The present study presentsthe changes in salivary immunoglobulins each month at thedifferent stations from the time of departure from Australiauntil isolation was broken at each station the following year.

Materials and Methods

Subjects

This study involved 16 expeditioners at Casey, 30 expeditioners atDavis and 27 expeditioners at Mawson participating in the 1992winter Australian National Antarctic Research Expeditions(ANARE). Three men at Davis and two men at Mawson wereexcluded from the study because they had only one or two samplescollected during the year. Overall, data were collected for 73 expe-ditioners, eight of whom were female. The median age of expedi-tioners was 31 years and was the same at each of the three stations.Ages ranged from 23 to 55 years (Table 1).

Saliva collections

Preliminary saliva samples were collected in Australia in lateNovember and early December 1991 and stored frozen at – 70°C atthe Australian Antarctic Headquarters in Hobart, Tasmania. Thestation doctors collected routine monthly saliva samples throughout1992. The samples were collected 2 h postprandially by gentle spit-ting into a collection tube. The saliva flow rate was not stimulated.At Casey and Davis, data were collected on 13 occasions (includingthe preliminary sample) and at Mawson on 12 occasions. The datacollection rate was very high, with Casey and Mawson collectingover 95% of all scheduled samples and Davis collecting 90%. Thesamples were stored immediately after collection at – 70°C andreturned to Australia frozen between December 1992 and January1993. All samples were stored frozen until assayed by a single tech-nician. The amount of data available for analysis after exclusions wasapproximately 90% of all collected samples (Table 1). Samples wereexcluded if they had deteriorated during storage, if there was insuf-ficient sample for analysis or if the subjects were fasting at the timeof collection. Fasting samples were excluded because the levels ofSIgA and SIgM are significantly elevated in fasting collection due toalterations in flow rate.9

Salivary immunoglobulins and albumin

Immunoglobulin A, IgG and IgM were measured in unstimulatedwhole mixed saliva by an in-house ELISA using commercially prepared unconjugated and biotin-conjugated antihuman antisera to detect IgA, IgG or IgM (Biosource International, Camarillo, CA,USA). Nunc maxisorp immunoplates (Nunc, Roskilde, Denmark)were coated with 100 µL of the unconjugated antisera 24 h prior toassay and stored at 4°C. The plates were returned to room tempera-ture (RT) and washed three times with a wash solution of phosphate-buffered saline (PBS, Dulbecco A; Oxoid, England) containing

0.05% Tween 20 (Aldrich, Steinheim, Germany), buffered to pH 7.2.The unbound sites on the plate were then blocked by adding 300 µLblocking buffer of PBS containing 5% skim milk (Diploma, Melbourne, Vic., Australia), to each well and incubating at RT for90 min. The plates were washed again three times to remove excessblocking buffer. An aliquot (100 µL) of each prediluted standard,control or saliva specimen was pipetted into appropriate wells andincubated at RT for 60 min. The plates were washed again prior topipetting 100 µL of the appropriate biotin-conjugated antisera (IgA,IgG or IgM) into each well and incubating at RT for 60 min. Theplates were washed to remove excess antisera prior to pipetting intoeach well 100 µL of streptavidin horseradish peroxidase (Dako,Glostrup, Denmark) solution, that had been prediluted 1:2000 inwash solution containing 0.1% skim milk. The plates were incubatedat RT for 30 min, washed three times and 150 µL substrate bufferadded to each well. The substrate buffer was prepared just prior touse by adding 10 µL of 3% hydrogen peroxide (Pharmacia, Uppsala,Sweden) to 10 mL ABTS substrate buffer, consisting of 2-2 Azino-bis-3 ethyl benzthiazoline 6 sulphuric acid (ABTS) tablets (Sigma,St Louis, MO, USA) dissolved in a citrate buffer at pH 4.5. The reac-tion was stopped at 15 min by adding 50 µL of a 1% solution of SDS(LKB, Bromma, Sweden). The absorbance of each well was read at405 nm using Molecular Devices Vmax Kinetic Microplate Reader(Molecular Dynamics, Sunnyvale, CA, USA). The standard curveand concentrations of samples were calculated using SOFTmax software (Molecular Dynamics). The assays were calibrated with Standard Human Serum referenced against BCR CRM470 Behring-werke, Marburg, Germany). Quality control was accomplished byincluding a range of known positive (high, medium and low) salivacontrols on all assays of the three classes of immunoglobulins. Inaddition, saliva from an IgA-deficient individual was included on allIgA assays as a negative control. Salivary albumin was measured byrate nephelometry using a Beckman ARRAY analyser, standards andcontrols (Beckman-Coulter, Brea, CA, USA).

Statistical analysis

The monthly salivary protein concentrations were compared at thedifferent stations and over the year using repeated measures analysisof variance with the Huynh–Feldt correction factor as implementedby the statistical package Stata.14 A repeated measures procedure wasused to properly account for the correlation of multiple measures onthe same individual. Because the distributions of the salivary proteinconcentrations were skewed, the logged values were used in theA N OVA procedures.

One-way analysis of variance was used to compare individualmean concentrations between stations for Fig. 2 using the statisticalpackage JMP.15 This was done to illustrate the station similarities anddifferences in the protein values indicated by the results of therepeated measures A N OVA . A mean was calculated for each individ-ual from all of their concentrations over the year and the distributionsof these mean values were compared between the stations. The

Mucosal immunity in Antarctica 617

Table 1 Demographic and sample collection details for the winter expeditioners at the Australian Antarctic stations in 1992

Station No. subjects Female:male ratio Median age (range; years) % Samples collected % Results after exclusions*

Casey 16 2:14 30.6 (23.7–42.1) 95.2 (198/208) 94.4–97.0Davis 30 5:25 31.4 (24.0–55.5) 90.3 (352/390) 88.4–93.2Mawson 27 1:26 31.7 (26.1–51.2) 95.4 (309/324) 88.3–90.0All 73 8:65 31.6 (23.7–55.5) 93.2 (859/922) 89.8–92.9

*Percentages differed for each protein if insufficient sample was available for analysis. Lowest and highest percentages are indicated.

Tukey–Kramer test was used for the multiple pairwise comparisonsof the stations, to indicate the nature of the station differences.

Person-centred values were calculated to illustrate the patterns ofchange in salivary proteins at each station over time for Fig. 3. Foreach expeditioner at each month, their yearly mean concentration wassubtracted from their monthly concentration to give the amount thattheir monthly concentration was above or below their individualyearly mean concentration at that month. The concentration differ-ences were then averaged for each month. This indicated when proteinconcentrations tended to be higher or lower for expeditioners duringthe year. Results of the repeated measures ANOVA indicated whetherconcentrations were different between the months (Table 3), but werenot useful for illustrating the specific nature of the differences.

Results

Differences in salivary proteins at each station

The monthly medians of salivary immunoglobulin andalbumin values measured in Tasmania prior to departure(month 0) and at the Antarctic stations of Casey, Davis andMawson in 1992 (months 1–12) are represented in Fig. 1. Themedian values and ranges of all predeparture samples and forsamples at each Antarctic station are indicated in Table 2.Results of repeated measures A N OVA assessing differencesbetween stations and the months of the study are presented in Table 3. The differences between stations tended to be

M Gleeson et al.618

Figure 1 Monthly median salivary IgA, IgM, IgG and albumin concentrations for the winter expeditioners at Casey (C), Davis (D) andMawson (M) from the time of departure from Australia in 1991 (month 0) and for each month spent in Antarctica during 1992.

Table 2 Median concentrations and ranges for salivary IgA, IgG, IgM and albumin in samples from 1992 winter expeditioners

Station Samples (n) IgA (mg/L) IgG (mg/L) IgM (mg/L) Albumin (mg/L)

PredepartureCasey 7 46.7 (36–96) 14.1 (98–32) 2.5 (1–4) 48.4 (15–271)Davis 13 41.8 (15–80) 9.2 (2–33) 2.2 (1–6) 40.3 (22–115)Mawson 23 59.9 (24–172) 29.1 (2–68) 2.5 (0–7) 54.1 (25–212)

AntarcticaCasey 185 57.1 (12–175) 15.1 (1–39) 4.0 (0–22) 47.4 (10–281)Davis 315 53.2 (12–237) 7.6 (1–49) 3.1 (1–27) 53.7 (16–252)Mawson 255 67.9 (22–216) 16.2 (2–52) 3.3 (0–30) 40.4 (9–252)

Data are shown for samples collected from the 1992 winter expeditioners prior to departing Australia in 1991 (predeparture) and for all monthsspent in Antarctica.

consistent over the year and are summarized in the graphs ofthe mean log salivary protein values for individuals at eachstation over the 12–13-month period in Fig. 2. Each point inFig. 2 represents an individual’s yearly mean value.

Immunoglobulin A There were significant differences inmean log(IgA) values between stations (P < 0.008; Table 3).On average, the log(IgA) values were higher at Mawson (geo-metric mean (gm) = 70.3 mg/L) than Casey (gm = 55.5 mg/L)and Davis (gm = 54.9 mg/L; Fig. 2). Within-person variabil-ity was not significantly different between the stations forlog(IgA) (P = 0.70).

Immunoglobulin G There were also significant differencesin station mean log(IgG) values (P < 0.0001; Table 3). Onaverage, the mean log(IgG) values at Davis (gm = 7.9 mg/L)were lower than those at Casey (gm = 14.3 mg/L) andMawson (gm = 15.9 mg/L; Fig. 2). Within-person variabilityfor log(IgG) was significantly different between the stations(P < 0.003). Casey tended to have the lowest within-personvariability and Davis the highest.

Immunoglobulin M There were no significant differences inmean log(IgM) values between stations (P = 0.85; Table 3).The geometric mean log(IgM) values for Casey, Davis andMawson were 3.7 mg/L, 3.3 mg/L and 3.4 mg/L, respectively(Fig. 2). Within-person variability for log(IgM) values wasdifferent between the stations (P < 0.003). Casey andMawson had similar within-person variability, while the variability at Davis was significantly lower.

Albumin There were no significant differences in stationmean log(albumin) values (P = 0.17; Table 3). Mawson(gm = 43.0 mg/L) tended to have lower values than Casey(gm = 51.1 mg/L) and Davis (gm = 55.9 mg/L; Fig. 2). Therewere also no significant differences in within-person variability of log(albumin) values between the stations(P = 0.91).

Mucosal immunity in Antarctica 619

Table 3 Results of repeated measures A N OVA : Significance levels(P values) for differences in protein concentrations between stations,months, and for interaction of stations and months

Salivary protein Station Month* Station–month interaction*

IgA 0.008 0.001 0.19IgG < 0.0001 0.08 0.28IgM 0.85 < 0.0001 0.006Albumin 0.17 0.37 0.18

*Huynh–Feldt corrected.

Figure 2 Box plots representing the distributions of the mean log salivary IgA, IgM, IgG and albumin concentrations for individuals atCasey (C), Davis (D) and Mawson (M) for the 1992 Antarctic winter expedition. The Tukey–Kramer plots provide a visual representationof the degree of overlap of the distributions of the data for each station.

Changes in salivary proteins during the year

There were significant differences between the monthly sali-vary protein values for IgA (P = 0.001) and IgM (P < 0.0001),but not for IgG (P = 0.08) or albumin (P = 0.37) monthlyvalues (Table 3). The stations tended to have similar patternsof change in immunoglobulin and albumin values over timein Antarctica, as illustrated with mean person-centred valuesin Fig. 3. There were distinct patterns to these monthlychanges for salivary IgA and IgM. No patterns were apparentfor salivary IgG or albumin.

Salivary IgA and IgM The significant monthly differencesfor salivary IgA and IgM followed a consistent pattern at allstations. Low values occurred in the first 4 months(January–April). The values increased to maximum levels inJuly and August and decreased back towards mean levels byDecember (Fig. 3).

Salivary IgG and albumin Changes in IgG values showedlittle variability over the 12 months of 1992 at all stations.Overall, monthly values of salivary IgG were not signifi-cantly different over the year (P = 0.08; Table 3). However,prior to departure from Australia (month 0) the meanmonthly differences were relatively higher for the Mawsonexpeditioners. Concentrations of IgG at Mawson were on

average 8.8 mg/L higher than individuals’ yearly mean IgGconcentrations, compared to Davis with concentrations2.1 mg/L higher and Casey with concentrations slightly lowerat –0.6 mg/L (Fig. 3). Albumin values did not change in anysystematic way over the study period and monthly valueswere not significantly different (Fig. 3; Table 3).

Discussion

The results of the present study revealed significant changesin salivary immunoglobulin patterns in the expeditioners overthe 12–13 months in Antarctica. The patterns during 1992were similar at the three Australian stations examined, withthe exception of higher IgG levels in the Mawson expedi-tioners prior to departing Australia. The concentrations ofsalivary immunoglobulins between subjects were signifi-cantly different for IgA and IgG, but showed no differencesfor IgM and albumin. The concentrations of IgA were higherat Mawson (P < 0.008) compared with the other stations andthe concentrations of IgG were lower at Davis (P < 0.001).

The causes of the statistically significant differences inconcentrations between the stations are unknown, but are relatively small differences in clinical terms, because allsamples had concentrations within the Australian populationreference ranges.16 The differences observed between stationsfor absolute concentrations of salivary immunoglobulins

M Gleeson et al.620

Figure 3 Changes in monthly mean person-centred IgA, IgM, IgG and albumin concentrations for the winter expeditioners at Casey(C), Davis (D) and Mawson (M) from the time of departure from Australia in 1991 (month 0) and for each month spent in Antarctica overthe 1992 Australian National Antarctic Research Expedition season.

(Fig. 1) were eliminated when adjusted to the person-centredmeans (Fig. 3), with the exception of the higher levels of IgGfor Mawson expeditioners prior to departure from Australia.Elevated levels of salivary IgG accompanied by high levels ofsalivary albumin usually reflect increased mucosal membranepermeability in the oral cavity11 and may be the cause of thehigher IgG levels observed in the Mawson expeditionersprior to departure from Australia (Figs 1,3). Disruption to theintegrity of the mucosa can be caused by many factors, suchas oral infections, dental disturbances or disruptions causinginflammation, but the causes in the present study areunknown.

The patterns of change in salivary IgA and IgM reflectedmucosal immune suppression in the first 4 months comparedwith the later part of the year in polar isolation. Of the factorsmost likely to be associated with this pattern of change inmucosal immunity in Antarctica, psychological stress wouldbe a significant contributor. Psychological stress and anxietyare well-recognized suppressors of salivary IgA.17–19 Theincreases in salivary IgA and IgM to mean levels began inApril, shortly after the commencement of the period of totalisolation for the winter expeditioners (March-November)when the summer expeditioners had departed. Studies on 18wintering groups, using techniques pioneered on smallremote Antarctic summer traverses,20 report that group dyna-mics appear to stabilize after the departure of the summerexpeditioners and the increase in salivary IgA may reflect aperiod of greater emotional stability at these stations in 1992(J Wood, pers. comm., 2000). Increases in salivary IgA havebeen reported to be associated with positive mood states21,22

and better coping skills23,24 in individuals and this would beconsistent with the observation of increased IgA and IgM inthe isolation period. Because psychological testing was notrecorded in the present study, any association betweenchanges in salivary IgA and psychological stressors can onlybe speculative. Increased anxiety has also been shown to havea negative effect on cell-mediated immunity in Antarcticexpeditioners over a summer season, with perceived anxietygreater before and after the voyage.2

Because the mucosal immune suppression occurred earlyin the expedition, and not during the months of colder tem-peratures and reduced daylight, any association with meteo-rological conditions is less likely. However, acclimatization tothe cold in Antarctica is reported to occur within the first fewmonths of arrival in Antarctica and post-cold-exposure stressdeclines over the period in Antarctica.25 Catecholamines, particularly noradrenaline, are associated with the adaptationto cold.26 The pattern of excretion of adrenaline for the winterexpeditioners at Casey in 197026 showed a remarkable simi-larity to the changes observed in salivary IgA and IgM at thethree stations studied in 1992. This adds weight to thehypothesis that the higher IgA levels (relative to the individ-ual) in the later months are associated with a positive adap-tive process.

The return to mean levels in the later period of the studycan also be explained by the changes in psychological stres-sors that occur in anticipation of the return to Australia.2,27

However, the reports of significant depressive symptoms inAntarctic expeditioners during the mid-winter isolation27,28

can not be ignored. Their association with changes inmucosal immunity warrants further investigation, because the

earlier report of an association between anxiety and measuresof systemic immunity2 was not verified in a more recentstudy.29

Another possible explanation for the relatively higherlevels of salivary IgA and IgM in the mid-winter monthswould be if the increases were associated with mucosal infec-tions. However, in this 1992 cohort the incidence of mucosalinfections was extremely low (19 infections in 16 of 73 expe-ditioners) and the majority of the infections (15/19) occurredprior to isolation of the winter expeditioners (ANAREHealth Register), consistent with previous reports of a lowincidence of infection in Antarctic winter expeditioners.25

Thus, mucosal infections were unlikely to be the cause of theincreased levels of salivary IgA and IgM in this cohort inmid-winter.

The present study presents the first data on mucosalimmune parameters in Antarctic expeditioners. The changesobserved over the Antarctic season were significant, but didnot support the original concern that severe immunosuppres-sion may occur in Antarctica. Further studies are required toidentify the causes of the changes in mucosal immunity inAntarctica and any implications for monitoring the health offuture Antarctic expeditioners.

Acknowledgements

The authors are grateful to the ANARE members who par-ticipated in this study during the winter of 1992 and to MrAdrian Flanagan (Hunter Immunology Unit) and Dr PeterSullivan (Polar Medicine, Australian Antarctic Division) fortheir assistance with collation of data. The authors acknowl-edge the assistance of Ms Sharron Hall, who assayed allsaliva samples, and Ms Jennifer Hutchings (Hunter Immunol-ogy Unit). The work was supported by the Antarctic ScienceAdvisory Committee Grants Scheme of The AustralianAntarctic Division and the study was approved by The EthicsCommittee of the Australian Antarctic Division.

References

1 Castrelos OD, Ikonicoff S, Zunino CB, Margni RA. Humanimmune humoral response in Antarctica. Instituto AntarcticoArgentino Contribucion 1983; 274: 1–26.

2 Muller HK, Lugg DJ, Ursin H, Quinn D, Donovan K. Immuneresponses during an Antarctic Summer. Pathology 1995; 27:186–90.

3 Williams DL, Climie A, Muller HK, Lugg DJ. Cell-mediatedimmunity in healthy adults in Antarctica and the subantarctic. J. Clin. Lab. Immunol. 1986; 20: 43–9.

4 Muller HK, Lugg DJ, Quinn D. Cell mediated immunity inAntarctic wintering personnel; 1984–92. Immunol. Cell Biol.1995; 73: 316–20.

5 Tingate TR, Lugg DJ, Muller HK, Stowe RP, Pierson DL.Antarctic isolation: Immune and viral studies. Immunol. CellBiol. 1997; 75: 275–83.

6 Holmes MG, Reed SE, Stott EJ, Tyrrell DAJ. Studies of experi-mental rhinovirus type 2 infections in polar isolation and inEngland. J. Hyg. Camb. 1976; 76: 379–93.

7 Cameron AS. Staphylococcal epidemiology in Antarctica. J. Hyg. Camb. 1970; 68: 43–52.

Mucosal immunity in Antarctica 621

M Gleeson et al.622

8 Brandtzaeg P, Baekkevold ES, Farstad IN et al. Regional specialisation in the mucosal immune system: What happens in the microcompartments? Immunol. Today 1999; 20: 141–51.

9 Brandtzaeg P, Jahnsen FL, Farstad IN. Immune functions andimmunopathology of the mucosa of the upper respiratory path-ways. Acta Otolaryngol. 1996; 116: 149–59.

10 Persson CG, Erjefalt JS, Erjefalt I et al. Contribution of plasma-derived molecules to mucosal immune defence, disease andrepair in the airways. Scand. J. Immunol. 1998; 47: 302–12.

11 Gleeson M, Cripps AW, Clancy RL. Modifiers of the humanmucosal immune system. Immunol. Cell Biol. 1995; 73: 397–404.

12 Lugg DJ. Current international human factors research inAntarctica. In: Harrison AA, Clearwater YA, McKay CP (eds).From Antarctica to Outer Space: Life in Isolation and Confine-ment. New York: Springer-Verlag, 1990; 31–42.

13 Mishra SK, Segal E, Gunter E et al. Stress, immunity andmycotic diseases. J. Med. Vet. Mycology 1994; 32 (S1): 379–406.

14 StataCorp. Stata Statistical Software: Release 6.0. CollegeStation (TX): Stata Corporation, 1999.

15 SAS Institute Inc. JMP, Version 3 Software. Cary (NC): SASInstitute Inc., 1999.

16 Gleeson M, Cripps AW, Clancy RL, Geraghty SB. The vari-ability of immunoglobulins and albumin in saliva of normal and IgA-deficient adults. In: MacDonald TT, Challacombe SJ,Bland PW, Stokes CR, Heatley RV, Mowat AMcI. (eds).Advances in Mucosal Immunology. Lancaster (UK): KluwerAcademic Press, 1990; 500–1.

17 Jemmot JB, Borysenko M, Chapman R et al. Academic stress,power motivation, and decrease in secretion rate of salivarysecretory immunoglobulin A. Lancet 1983; i: 1400–2.

18 Graham NMH, Bartholomeusz RCA, Taboonpong N, La BrooyJT. Does anxiety reduce the secretion rate of secretory IgA insaliva? Med. J. Aust. 1988; 148: 131–3.

19 Ng V, Koh D, Chan G, Ong HY, Chai SE, Ong CN. Are salivaryimmunoglobulin A and lysozyme biomarkers of stress amongnurses? J. Occup. Environ. Med. 1999; 41: 920–7.

20 Wood J, Lugg DJ, Hysong SJ, Harm DL. Psychological changesin hundred-day remote Antarctic field groups. Environ. Behav.1999; 31: 299–337.

21 Kubitz KA, Peavy BS, Moore BS. The effect of daily hassles ofhumoral immunity: An interaction moderated by locus ofcontrol. Biofeedback Self Regul. 1986; 11: 115–23.

22 Stone AA, Valdimarsdottir H, Jandorf L, Cox DS, Neale JM.Evidence that secretory IgA antibody is associated with dailymood. J. Pers. Soc. Psychol. 1987; 52: 988–93.

23 Dillon KM, Minchoff B, Baker KH. Positive emotional statesand enhancement of the immune system. Int. J. Psychiatry Med.1985; 15: 12–17.

24 Henningsen GM, Hurrell JJ, Baker F et al. Measurement of salivary immunoglobulin A as an immunologic biomarker of job stress. Scand. J. Work Environ. Health 1992; 18 (S2): 133–6.

25 Lugg DJ. Antarctic epidemiology: A survey of ANARE Stations1947–72. In: Edholm OG, Gunderson EKE (eds). Polar HumanBiology. London: Heinemann, 1973; 93–104.

26 Bodey AS. The role of catecholamines in human acclimatizationto cold: A study of 24 men at Casey, Antarctica. In: Edholm OG,Gunderson EKE (eds). Polar Human Biology. London: Heine-mann, 1973; 141–9.

27 Godwin JR. A preliminary investigation into stress in AustralianAntarctic expeditioners. SPRIO Polar Symposia 1991; 1: 9–22.

28 Palinkas LA, Cravalho M, Browner D. Seasonal variation ofdepressive symptoms in Antarctica. Acta Psychiatr. Scand.1995; 91: 423–9.

29 Pitson GA, Lugg DJ, Muller HK. Seasonal cutaneous immuneresponses in an Antarctic wintering group: No association withtestosterone, vitamin D metabolite or anxiety score. Arctic Med.Res. 1996; 55: 118–22.