Ventilation rate buildings building syndrome

Occupational and Environmental Medicine 1995;52:709-714

Ventilation rate in office buildings and sickbuilding syndrome

Jouni JK Jaakkola, Pauli Miettinen

AbstractObjective-To examine the relationbetween ventilation rate and occurrenceof symptoms of the eyes, nose, throat,and skin as well as general symptomssuch as lethargy and headache, oftentermed the sick building syndrome.Methods-A cross sectional populationbased study was carried out in 399 work-ers from 14 mechanically ventilated officebuildings without air recirculation orhumidification, selected randomly fromthe Helsinki metropolitan area. The ven-tilation type and other characteristics ofthese buildings were recorded on a sitevisit and the ventilation in the roomswas assessed by measuring the airflowthrough the exhaust air outlets in theroom. A questionnaire directed at work-ers inquired about the symptoms andperceived air quality and their possiblepersonal and environmental determi-nants (response rate 81%). The outcomeswere weekly work related symptoms expe-rienced during the previous 12 monthsand symptom groups defined either bytheir anatomical location or hypothesisedmechanism.Results-In logistic regression analysis,the adjusted odds ratio (OR) for anysymptom of interest was 3 03 (95% confi-dence interval (95% CI) 1-13 to 8-10) in thevery low ventilaton category ofbelow 5 Usper person and 2-24 (0-89 to 5-65) in thehigh ventilation category of over 25 Us perperson compared with the reference(15-<25 Us). The ORs for ocular (1.27,1 11 to 1-46), nasal (1.17, 1-06 to 1.29), skinsymptoms (1*18, 1.05 to 1-32), andlethargy (1.09, 1-00 to 1.19) increased sig-nificantly by a unit decrease in ventilationfrom 25 to 0 Us per person.Conclusion-The results suggest thatoutdoor air ventilation rates below theoptimal (15 to 25 Us per person) increasethe risk of the symptoms of sick buildingsyndrome with the sources of pollutantspresent in mechanically ventilated officebuildings. The Finnish guideline value is10 Us per person.

(Occup Environ Med 1995;52:709-714)

Keywords: building ventilation; air pollution, indoor;sick building syndrome

Office workers commonly report the indoorair quality to be unacceptable and experience

non-specific symptoms of the eyes, nose,throat, and skin as well as general symptomssuch as lethargy and headache.'-8 The similarityof the symptom pattern in different study pop-ulations and their causal attribution to thebuilding environment has given rise to theconcept of the sick building syndrome.9

Air change plays a central role in loweringthe levels of indoor air pollutants emitted fromthe indoor environment or produced by theoccupants or their activities such as smoking,photocopying, or handling paper. It is com-monly accepted that ventilation is necessaryfor the wellbeing of the occupants. The roleof ventilation is complicated by the observa-tions that workers in air conditioned build-ings experience more symptoms thanworkers in naturally ventilated buildings.7 10 1Mechanical ventilation-that is, mechanicalsupply and exhaust of air-has also beenrelated to these symptoms in some studies,67 10-1but not in others.'3 14 Six earlier studies, fiveexperimental'5"'9 and one observational420have assessed the relation between ventilationrate and occurrence of symptoms. The resultsare inconsistent, although the weight of evi-dence has been interpreted to suggest thatventilation rates below 10 1/s per personwould increase the risk of symptoms ofsick building syndrome.2' These results donot provide sufficient information to deter-mine the optimum outdoor ventilation rateneeded in the existing buildings to protect theoccupants from adverse health effects.We examined the relation between ven-

tilation rate and occurrence of symptoms ofthe sick building syndrome in office workersin a population based cross sectional studycarried out in a Helsinki metropolitan area.

MethodsBUILDINGS AND STUDY POPULATIONA random sample of 50 office buildings with atleast 10 workers in each was selected from theHelsinki Metropolitan Area Building Registry,which included all the 894 office buildings inthe defined geographical region.'0 The ownersof nine buildings were not willing to partici-pate in the study. The remaining 41 buildingsand their occupants were included in thestudy, and a person registry was made of allthe workers. In six buildings a few firms, com-prising only a small proportion of the workers,did not participate. A self administered ques-tionnaire was posted to 3317 workers. A totalof 2678 workers completed the questionnaire(response rate 81%), 605 subjects did not par-ticipate, and 34 returned an incomplete form.

EnvironmentalEpidemiology Unit,Department of PublicHealth, University ofHelsinki, FinlandJouni JK JaakkolaDepartment ofEpidemiology,National Institute ofPublic Health, Oslo,NorwayJouni JK JaakkolaLaboratory of Heating,Ventilating and Air-Conditioning, HelsinkiUniversity ofTechnology, Espoo,FinlandJouni JK JaakkolaPauli MiettinenCorrespondence to: DrJaakkola, Department ofPublic Health, PO Box 21,FIN-00014 University ofHelsinki, Finland.Accepted 29 June 1995

709

on October 23, 2021 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.52.11.709 on 1 N

ovember 1995. D

ownloaded from

http://oem.bmj.com/

70aakkola, Miettinen

Table 1 Characteristics of the study population

Ventilation category (l/s per person)

Characteristic

Personal characteristicsSex:

FemaleAge:

< 2425-3435-4445-54>55

Professional education:University or collegeVocational schoolNoOtherAtopy

SmokingCurrentFormer

Office environment, work:Shared officePassive smokingTextile wall materialsWall to wall carpetsSealed windowsHandling of

self copying paperVDT use (h/week,mean (SD))

Photocopying (h/week,mean (SD))

Psychosocial environment:Atmosphere:

Satisfied or neutralUnsatisfiedVery unsatisfied

Stress:Slight, not at allModerateMuch, very much

Interest in work:Often, almost alwaysSometimesRarely, never

Very low(< 5)n = 53(%o)

73-6

3-828-335.915-117-0

60-416-714-68-3

45-3

30-220-8

64-25-7

11-31-9

15-124-5

19-1(12-4)1-7

(1-8)

86-511-51-9

45-332-122-6

67-926-45-7

Low(5-<15)n = 171(%/-)

56-1

11 736-328-720-52-9

64-911-316-17.7

29-8

32 218 7

6430-6

13-55.99.4

26-9

19-4( 1-9)2-2

(6-0)

87-19.43.5

43-336-819-9

67-225-77 0

Medium(15-<25)n = 123(Go)

43.94.9

28-531-726-88 1

80-58-15-75-7

37-4

34-221-1

38-21 68-2

17-99-8

29-3

18-8(12-2)1-8

(2-2)

84-614-60-8

45-539015 5

70 724-44-9

High() 25)n = 52(%O)

51-9

11-517-334-632-73.9

66-018-08-08-0

32-7

32-719-2

32-70.0

32-719-215-430-8

16-2(9-5)2-6

(6-1)

88-57-83.9

38-548-113-5

78-819-21-9

54-1

8-530 331-323-36-5

69-411-811-37-0

34-6

32-619-8

52-11-5

14-010-811-027-8

18-8(11-8)2-1

(4-7)

86-211-12-5

43-638-518-0

69-924-65-5

The present study focused on the mechani-cally ventilated buildings (both supply andexhaust) that were operated without air recir-culation or humidification. A total of 14 build-ings, were included; 10 buildings were fittedwith simple mechanical ventilation and fourwere air conditioned. Simple mechanical ven-tilation refers to buildings with ducted supplyand exhaust airflow, but no integrated heatingor cooling of air. Partial heating of incomingair is included in this category. Buildings withair conditioning but without humidificationare buildings with integrated heating andcooling of incoming air. The study population(table 1) consisted of 399 subjects, 183 men

(45.9%) and 216 women (54.1%).

OUTCOME CRITERIAThe outcomes were nine particular symptomscompatible with those described in the sickbuilding syndrome,9 including: eye symptoms;nasal dryness, itching or irritation; nasal con-gestion (stuffy, blocked nose); nasal excretion(runny nose); pharyngeal symptoms; skinsymptoms; headache; lethargy; and difficultyin concentrating. Apart from particular symp-toms, summary outcomes were also consid-ered, based on either the anatomical locationof the symptoms (ocular, nasal, pharyngeal,cutaneous, or general) or on the hypothesisedmechanism (mucosal irritation or allergicreaction).

The outcomes were specified as appearingat least weekly, being work related, and thereference time period was the previousyear. The questions used are described indetail elsewhere.'0 These weekly work relatedsymptoms were used in forming the summaryoutcomes. The presence of one or more symp-toms in the corresponding symptom groupindicated a positive response. Mucosal irritationwas based on reported dryness, itching or irri-tation of eyes; nasal dryness; nasal congestion;or pharyngeal irritation. Allergic reaction wasbased on reported dryness, itching, or irrita-tion of eyes; nasal congestion; nasal excretionor pharyngeal irritation.

DATA COLLECTIONA self administered questionnaire askedabout the symptoms typical of the sick build-ing syndrome and perceptions of odour andthermal comfort. Personal factors, informationon health, smoking, and work environmentwere also asked for in the questionnaire. Thequestionnaire was distributed by post andrequested twice from the non-responders.

Basic information on the building, includ-ing the size and type of building, was availablein the Building Registry. Further informationwas collected in an interview with the land-lords and custodians of the buildings and onsite visits.The air change in the rooms was assessed

from November 1991 to February 1992 bymeasuring the airflow through the exhaustair outlets in the room. An anemometer tubeconnected to a hot wire anemometer was usedin the measurements. The anemometer wasregularly calibrated in the laboratory, no driftin the calibration of the anemometer wasdiscovered during the study. This methodgives a good estimate of the total air change ina room, as all mechanically ventilated officebuildings in Finland are designed to havenegative pressure. The floor area and volumeof the rooms were measured at the same timeas the ventilation measurements. From thesefigures, the ventilation rates per person in eachroom were calculated.

STATISTICAL METHODSThe age (five categories) and sex standardizedprevalences of weekly, work-related symptomswere estimated in four different ventilationcategories: very low (< 5 1/s per person), low(5-< 15), medium (15-<25) and high (>25 1/sper person). The relation between the level ofairflow and occurrence of symptoms was thenassessed in logistic regression analysis.22 23Adjusted odds ratios (ORs) were computedfor the very low, low, and high ventilationcategories by the indicator technique with themedium ventilation category as a control,because of the lowest prevalences.22 The fol-lowing covariates were included in the models:age, sex, professional education, atopy (anyhistory of eczema, allergic rhinitis, conjunctivi-tis, or asthma), personal smoking, type ofventilation (mechanical or air conditioned),number of workers in the room, passive smok-ing, textile wall materials, wall to wall carpets,

710



.bmj.com

/O

ccup Environ M


ovember 1995. D

ownloaded from

http://oem.bmj.com/

Ventilation rate in office buildings and sick building syndrome

Table 2 Age and sex standardised prevalences (%) of weekly, work related symptomsduring previous 12 months in different ventilation categories

Ventilation category (i/s per person)

Very low Low Medium High(<5) (5-<15) (15-<25) (, 25) Total

n = 53 n = 171 n = 123 n = 52 n = 399(%) (C/) (%/0) (%/0) (0,0)

Eye symptoms 15-8 5-7 4-3 17-6 8-5Nasal irritation 11-2 12-8 5-0 7-7 10-0Nasal congestion 14-3 4-5 2-2 9-6 6-0Nasal discharge 13-1 3-9 1-7 5-7 5-3Pharyngeal symptoms 8-4 7-4 0 0 7-7 4-5Skin symptoms 13-0 5-1 0 9 9 7 5 0Headache 53 6-3 3-4 6-8 5-0Lethargy 14-6 12-6 10-4 13-5 11-5Difficulty concentrating 4-6 8-1 5-9 10-8 6-3

sealed windows, work with a video displayunit, handling of self copying paper, photo-copying, satisfaction with psychosocial atmos-phere, work stress, interesting work, average

weekly time in the office, and duration ofemployment in the building. Finally, paramet-ric models were fitted, when applicable, toexamine the form of the relation between in

odds of the outcome and ventilation rate. Forsome outcomes a linear relation was foundbetween In odds and ventilation rate in therange of 0 to 25 1/s per person. This corre-

sponds to an exponential relation betweenthe risk of symptoms and ventilation rate inthis range. When there were problems withconvergence of the full model, a forwardstepwise modelling approach was applied withventilation rate, sex, and age forced into themodel and other variates selected according tothe improvement of fit (0O25 significance levelas the criterion for entry and exclusion).2'

ResultsSTUDY POPULATIONThe distributions of the potential confoundersvaried to some extent among the ventilationcategories (table 1). The proportion ofwomen was considerably greater in the very

low (73-5%) and low ventilation categories(56 1%) than in the medium (43-9%) and

high ventilation categories (51 9%), and theage distribution varied among the ventilationcategories. Atopy, shared office, passivesmoking, sealed windows, and work stresswere more common in the low ventilationcategories, whereas textile wall materials, wallto wall carpets, and handling of self copyingpaper were most common in the high ventila-tion category.

VENTILATION RATE AND SYMPTOMSThe median ventilation rate in the rooms was

13-8 1/s per person, 106 1/s per person forbuildings with simple mechanical ventilationand 16 5 1/s per person for air conditionedbuildings. The proportion of subjects whooccupied offices with the ventilation ratebelow 5 1/s per person was 13%, and 56% ofthe subjects were in roons with the ventilationrate below 15 1/s per person.The age and sex standardised prevalences of

weekly work related symptoms were the lowestin the medium ventilation category for all thesymptoms except for difficulty in concentrat-ing (table 2). Ocular, nasal, pharyngeal andskin symptoms and lethargy showed an

increasing trend from the medium to very lowventilation category. The prevalences of symp-toms in the high ventilation category were alsoconsiderably greater than in the medium ven-

tilation category.The adjusted risk estimates provided similar

evidence of the relation between ventilationrate and risk of symptoms (table 3). The riskestimates in the lowest ventilation categorywere highest for skin (OR 41-0), nasal (7 7),and eye symptoms (5-6), but also considerablyincreased for general symptoms (2 1) whencompared with the medium ventilation cate-gory. All except general symptoms were signif-icant. The corresponding risk estimates in thehigh ventilation category were also all raised,ranging from 1-1 to 14-8, significant for eye

symptoms. The adjusted ORs for mucosalirritation (OR 4-21, 95% confidence interval(95% CI) 1-35 to 13-13) and allergic rea-

ction (7 04, 2-02 to 24 54) were significantlyincreased.

Table 3 Adjusted ORs of weekly, work related symptoms during previous 12 months in very low, low, and highventilation categories compared the with medium ventilation category (n = 399; ventilation rates are in I/s per person)

Very low (<5) Low (5-<15) Medium (15-<25) High (>25)

OR* OR* Reference OR*

Point Point Pointestimate 95% CI estimate 95% CI estimate 95% CI

Eye symptoms 5-57 (1-16-26-79) 2-31 (0 58-9 29) 1-00 4-32 (1-01-18-59)Nasal symptoms 7-66 (2-01-29-23) 5 90 (1-91-18-22) 1-00 2 15 (0 54-8 58)

Irritation 3-65 (0-80-16-59) 6-07 (1-78-20-71) 1-00 2-23 (0-48-10-26)Congestion 19-20 (3-04-121-1) 2-83 (0-56-14-35) 1-00 3-15 (0-53-18-69)Discharge 14-81 (1-72-127-7) 3-46 (0-46-26-2) 1-00 5-84 (0-61-55-75)

Pharyngeal symptoms - - 1-00 -

Skin symptoms 40-97 (2-15-781 2) 8-93 (0-63-126-0) 1-00 14 75 (0-92-237-5)General symptoms 2-07 (0 68-6 26) 1-96 (0-85-4 57) 1-00 1-63 (0 56-4-76)Headache 4-63 (0 66-32 38) 3-43 (0-63-18-68) 1-00 5-58 (0-84-37-14)Lethargy 2-46 (0 73-8-32) 1-44 (0 54-3-88) 1-00 1-12 (0-31-4-09)Difficulty concentrating 2-67 (0-39-18-20) 2-66 (0-69-10-30) 1-00 1-62 (0-27-9-81)

Any symptom 3 03 (1-13-8-10) 3-34 (162-6-91) 1-00 2-24 (0 89-5 65)Mucosal irritation 5-26 (1-57-17-60) 5 30 (1-97-14-23) 1-00 4-21 (1-35-13-13)Allergic reaction 8-96 (2-34-34-29) 5-24 (1-69-16-24) 1-00 7 04 (2 02-24 54)

*Adjustment was made in the logistic regression for age, sex, professional education, atopy (any history of eczema, allergic rhinitis,conjunctivitis, or asthma), number ofworkers in the room, personal and passive smoking, type of ventilation, textile wall materials,wall to wall carpets, sealed windows, work with a video display unit, handling of self copying paper, photocopying and satisfactionwith psychosocial atmosphere, job stress, interesting job, average weekly time in the office, and duration of employment in thebuilding.tNo subjects with symptoms in the control group.

711



.bmj.com

/O

ccup Environ M


ovember 1995. D

ownloaded from

http://oem.bmj.com/

7Jaakkola, Miettinen

Table 4 Increase in ORs of weekly, work relatedsymptoms during the previous 12 months per 1 1/s decreasein ventilation rate between 25 to O is per person (n = 347)

OR*

Point estimate 95% CI

Eye symptoms 1-27 1l 1-1 46tNasal symptoms 1-17 1-06-1-29

Irritation 1-10 1-00-1-22Congestion 1 30 1-13-1-50tDischarge 1-17 1-05-1-32t

Pharyngeal symptoms 1-12 0-98-1-27Skin symptoms 1-18 1-05-1-32tGeneral symptoms 1-05 0-97-1-14

Headache 1-01 0-89-1-15Lethargy 1-09 1 00-1 19Difficulty concentrating 1-02 0-90-1-15

Any symptom 1-09 1-02-1-17Mucosal irritation 1-15 1-05-1-25Allergic reaction 1-18 1-07-1-31

*As for table 3.tStepwise modelling with age and sex forced into the model.

Logistic regression models indicated asignificant linear increase in the ln odds ofeye, nasal, and skin symptoms and lethargy bya unit decrease in the ventilation rate from25 1/s towards 0 (table 4). The risk estimatesare expressed as a change in OR per 1 1/sdecrease in the ventilation rate.

DiscussionIn the present population based study, the riskof work related symptoms, typical of the sickbuilding syndrome, was found to be related tolow outdoor air ventilation rate in the office.Ocular, nasal, pharyngeal, and skin symptoms,as well as lethargy were considerably morecommon in the rooms with the ventilation ratebelow 5 1/s per person compared with therooms where the rate was considered to beoptimal, from 15 to 25 1/s per person. The riskof these symptoms was estimated to increaseexponentially by a unit decrease below theoptimal ventilation rate. The risk of symptomswas also increased in the rooms with the venti-lation rate above the optimum.

RELATION BETWEEN VENTILATION RATE ANDSICK BUILDING SYNDROMEThe relation between the ventilation rate andthe occurrence of symptoms is complicated,because it depends on several phenomena.Exposure to several chemical and biologicalindoor air pollutants can have adverse healtheffects. Air change is essential in lowering theair pollution levels and thus ventilation ratesthat are too low can be related to adversehealth effects when exposure to these pollu-tants is increased. High ventilation rates couldalso be related to adverse health effects in twoways. Firstly, outdoor air and the ventilationsystem are sources of indoor air pollution tothe extent that an increase in ventilation rateincreases the exposure. Secondly, mechanicalventilation changes the physical properties ofindoor air. Berglund and Lindvall suggestedthat the human sensory system uses a patternrecognition mechanism in the sensing ofindoor air with complex environmentaladaptation to the inhaled air.24 Use of high

ventilation rates can lead to extremehomogenisation of air, causing sensory con-fusion and strain on the system when it is tryingto interpret the signals. A possible hypothesisis that sensory symptoms tied to sick buildingsof the irritant type 24 might be related to thehomogenisation of air.

VALIDITY OF RESULTSThe potential sources of systematic error inthe assessment of the relation between airchange and occurrence of symptoms includeselection bias, information bias, and con-founding. There are two possible ways thatselection bias could have been introduced onan individual level. On the one hand, individualworkers with work related dissatisfaction andsymptoms could have left work. Presumingthat either low or high air change is a determi-nant of the symptoms, this type of selectionwould lead to underestimation of the relationbetween the air change and occurrence ofsymptoms. On the other hand, the risk esti-mates could have been overestimated, forexample if workers experiencing symptomsand perceptions were more likely to completethe questionnaire and the response rate wasrelated to low ventilation rate. This type ofselection was not likely, because the responserate was high (81%). An excess of symptomsin the high ventilation category compared withthe medium ventilation category could beexplained by a selective adjustment to increasethe ventilation rate as a response to occupantcomplaints. The time sequence between expo-sure and outcome can not be assessed in across sectional design, and thus the resultsshould be interpreted cautiously.The participants were informed about the

general objective of studying the environmen-tal determinants of health and well being inoffice buildings, but no special attention waspaid to the role of ventilation. Further, theparticipants did not know the magnitude ofventilation rate in their room. Thus, informa-tion bias is an unlikely explanation of theresults.The results, which indicated that the preva-

lence of symptoms is related to the magnitudeof ventilation rate, could be explained byuneven distributions of other determinants ofthe outcomes. Most of the known confounderswere taken into account in the multivariateanalysis. Other unknown confounders notconsidered in the analysis could theoreticallyexplain our results, but these would have to beboth strong determinants of the outcomes andbe associated strongly with the ventilation rate.Some random error was likely both in the

measurement of ventilation rates and out-comes. The airflows were measured about ayear after participating in the survey. Therooms, where renovation or adjustment ofventilation had taken place after the question-naire, were excluded from the study. Somechanges may have taken place in the buildingswithout our knowledge. The non-differentialmisclassification of both the ventilation rateand the outcomes would have led to underesti-mation of their relation.

712



.bmj.com

/O

ccup Environ M


ovember 1995. D

ownloaded from

http://oem.bmj.com/

Ventilation rate in office buildings and sick building syndrome

SYNTHESIS WITH PREVIOUS KNOWLEDGEThree earlier studies that assessed the relationbetween ventilation rate and occurrence ofsymptoms'41518 20 reported a significant associa-tion between low ventilation rates and theoccurrence of symptoms, and three studies'61719did not.Of the five experimental studies, only two

found an association between ventilation rateand symptoms. These studies provided strongdesigns for causal inference in the particularbuildings selected. A major limitation from thepublic health point of view was that the studieswere only carried out in between one and fourbuildings and thus the results can not be usedto evaluate the relation in the existing build-ings given the current sources of indoor airpollution. In the Finnish controlled parallelgroup trial the average symptom score for sickbuilding syndrome increased when the out-door ventilation rate was reduced from an

average of 26 1/s per person to below 7-5 1/s perperson.'5 In the American two part uncon-

trolled trial,'8 building occupants experiencedmore symptoms at lower than at higher out-door ventilation rates. The contrasts were 7-81/s and 17-9 1/s per person during the summerand 5-8 1/s and 14-9 1/s per person during thewinter. In the Swedish controlled trial,'7 theventilation rate was altered between an aver-

age 12 1/s and 23 1/s per person, but the alter-ation had no effect on the symptoms or signs,such as blink rates of the eyes, tear film stabil-ity, and visible signs of the eyes, lips, or skin.In the Finnish four period crossover trial'9 nodifference was found in the occurrence ofsymptoms between the average outdoor venti-lation rates of 6 1/s and 20 1/s per person. Nordid the similar Canadian six period crossover

trial'6 report any difference in the occurrence

of symptoms between the two compared levelsof airflows, 14 1/s and 30 1/s per person.The limitations of non-experimental

studies, whether cross sectional or case-

control studies, include the problems of biasand confounding, but a strength from thepublic health point of view is the feasibility ofstudying several buildings with populationbased sampling. This was applied both in thepresent cross sectional study and in the north-ern Swedish office illness project with a cross

sectional'4 and case-control study.20 In agree-ment with the present study, the Swedishstudy reported an increased risk of symptomsin the low ventilation category, below 8-5 1/sper person (OR 1-87, 95% CI 1.02 to 3 4) andin the medium ventilation category from 8 5 to13-6 1/s per person (1-50, 0-84 to 2 7), com-

pared with the control category of 13-6 1/s perperson.The range of outdoor air ventilation rate

considered in the studies is an important issue.Based on his review of the six earlier studies,Mendell suggested that ventilation rates at or

below 10 1/s per person were consistently asso-

ciated with an increased risk of symptoms,2'with the Finnish crossover trial' as an excep-tion. The present study is consistent with thissuggestion for the role of low ventilation rates.It should be emphasised that as well as the

strength of elimination, the strength of emis-sions also determines the level of exposure andconsequent health effects. The results of thepresent study suggest that in office buildingswith current sources of pollution, the risk ofsymptoms is already increased when out-door air ventilation rates are below 15 1/s perperson.

CONCLUDING REMARKSThe amount of ventilation needed in a givenspace depends essentially on the sources ofpollution in the indoor environment. Underthe most simple conditions, the subject him orherself is the only source of air pollution. Inmodem office environments, pollution sourcesalso include the usual human activities such assmoking, handling of paper, photocopying,and physical objects such as wall and floor sur-faces, furniture, and textiles. The situation iseven more complicated when the air is recircu-lated or the supply air is polluted.

Ideally, the interior materials should notemit pollutants. The activities worsening theair quality, such as smoking and photocopy-ing, should take place in separate rooms, andcleaning should be frequent enough to reducethe amount of accumulating particles on thesurfaces. Although increasing the air changeshould be a secondary precaution, the mini-mum ventilation requirements should beadjusted to the existing conditions in the officeenvironment. The present study was popula-tion based and thus provides information onthe ventilation requirements in the existingFinnish buildings. It seems that an outdoor airventilation rate below 5 1/s per person is astrong determinant of ocular, nasal, pharyn-geal, and skin symptoms, as well as generalsymptoms, and ventilation rates up to 15 1/sper person are likely to increase the risk ofthe symptoms. The Finnish guideline value is10 1/s per person.The present study provides new evidence of

an association between high ventilation ratesand the risk of symptoms. This could indicatethat mechanical ventilation systems may affectthe indoor air quality by changing the physicalproperty of air or by being a source of indoorair pollution. An alternative explanation is thathigh ventilation rates indicate a reaction to anexisting indoor air problem. This observationshould be interpreted with caution, because ofthe limitations of the study design.

This study was supported by the Ministry of Financial Affairs,Finland.

1 Finnegan MJ, Pickering CAC, Burge PS. Sick building syn-drome: prevalence studies. BMJ 1984;289:1573-5.

2 Jaakkola JJK. Indoor air in office building and human health.Experimental and epidemiologic study of the effects of mechan-ical ventilation. Helsinki: Department of Public Health,University of Helsinki, 1986. (Health Services Researchby the National Board of Health in Finland, No 41. InFinnish with an English summary. Doctoral thesis.)

2a Jaakkola JJK, Heinonen OP, Seppanen 0. Mechanicalventilation in office buildings and the sick buildingsyndrome. An experimental and epidemiological study.IndoorAir 1991;1:111-21.

3 Skov P, Valbj0rn 0, and DISG. The "sick" building syn-drome in the office environment: the Danish Town Hallstudy. Environ Int 1987;13:339-49.

4 Burge S, Hedge A, Wilson S, Harris-Bass J, Robertson A.Sick building syndrome: a study of 4373 office workers.Ann Occup Hyg 1987;31:493-504.

713



.bmj.com

/O

ccup Environ M


ovember 1995. D

ownloaded from

http://oem.bmj.com/

74aakkola, Mietmnen

5 Norback D, Rand G, Michel I, Amcoff S. The prevalenceof symptoms associated with sick buildings and pollutedindustrial environments as compared to unexposed refer-ence groups without expressed dissatisfaction. EnvironInt 1989;15:85-94.

6 Zweers T, Preller L, Brunekreef B, Boleij JSM. Health andindoor climate complaints of 7043 office workers in 61buildings in the Netherlands. IndoorAir 1992;2:127-36.

7 Fisk WJ, Mendell MJ, Daisey JM, Faulkner D, HodgsonAT, Nematollah M, Macher JM. Phase 1 of theCalifornia healthy building study: a summary. Indoor Air1993;3:246-54.

8 Stenberg B, Mild KH, Sandstrom M, Sundell J, Wall S. Aprevalence study of the sick building syndrome and facialskin symptoms in office workers. Indoor Air 1993;3:71-81.

9 Indoor air pollutants: exposure and health effects. Report ona World Health Organisation meeting. Copenhagen,Denmark: World Health Organisation Regional Officefor Europe, 1983. (EURO reports and studies No 78.)

10 Jaakkola JJK, Miettinen P. Type of ventilation system inoffice buildings and sick building syndrome. Am JEpidemiol 1995;141:755-65.

11 Skov P, Valbj0rn 0, Pedersen BV. Influence of indoor cli-mate on the sick building syndrome. Scand J WorkEnviron Health 1990;16:363-71.

12 Hedge A, Burge PS, Robertson AS, Wilson S, Harris-Bass J.Work-related illness in offices: a proposed model of the"sick building syndrome". Environ Int 1989;15: 143-58.

13 Mendell MJ, Smith AH. Consistent pattern of elevatedsymptoms in air-conditioned office buildings: a reanalysisof epidemiologic studies. Am J Public Health 1990;80:1193-9.

14 Sundell J, Lindvall T, Stenberg B. Associations betweentype of ventilation and air flow rates in office buildingsand the risk of SBS-symptoms among occupants. EnvironInt 1994;20:239-51.

15 Jaakkola jJK, Heinonen OP, Seppanen 0. Mechanicalventilation in office buildings and the sick buildingsyndrome. An experimental and epidemiological study.IndoorAir 1991;1:111-21.

16 Menzies R, Tamblyn R, Farant J-P, Hanley J, Nunes F,Tamblyn R. The effect of varying levels of outdoor-airsupply on the symptoms of sick building syndrome.NEnglJMed 1993;328:821-7.

17 Wyon D. Sick buildings and the experimental approach.Environmental Technology 1992;13:313-22.

18 Nagda NL, Koontz MD, Albrecht RJ. Effect of ventilationrate in a healthy building. In: Healthy buildings: Pro-ceedings of the ASHRAE IAQ'91 Conference. Washington,DC: ASHRAE, 1991;101-7.

19 Jaakkola JJK, Tuomaala P, Seppanen 0. Air recirculationand sick building syndrome. A four period crossover trial.Am J Public Health 1994;84:422-8.

20 Stenberg B. Office illness. The worker, the work and theworkplace. Umea, Sweden; 1994. ((UmeA UniversityMedical Dissertaion). New Series No 399.)

21 Mendell MJ. Non-specific symptoms in office workers: areview and summary of the epidemiologic literature.IndoorAir 1993;3:227-36.

22 Miettinen OS. Theoretical epidemiology. Principles of occur-rence research in medicine. New York: John Wiley, 1985.

23 Hosmer DW, Lemeshow S. Applied logistic regression. JohnWiley, 1989.

24 Berglund B, Lindvall T. Sensory reactions to "sick build-ings". Environ Int 1986;12:147-59.

Instructions to authorsThree copies of all submissions should besent to: The Editor, Occupational andEnvironmental Medicine, BMJ PublishingGroup, BMA House, Tavistock Square,London WC1H 9JR, UK. All authorsshould sign the covering letter as evidenceof consent to publication. Papers reportingresults of studies on human subjects must beaccompanied by a statement that thesubjects gave written, informed consentand by evidence of approval from theappropriate ethics committee. These papersshould conform to the principles outlinedin the Declaration of Helsinki (BMJ 1964;ii: 177).

If requested, authors shall produce thedata on which the manuscript is based, forexamination by the Editor.Authors are asked to submit with

their manuscript the names andaddresses of three people who theyconsider would be suitable independentreviewers. They will not necessarily beapproached to review the paper.Papers should include a structured

abstract of not more than 300 words,under headings of Objectives, Methods,Results, and Conclusions. Pleaseinclude up to three keywords or keyterms to assist with indexing.

714



.bmj.com

/O

ccup Environ M


ovember 1995. D

ownloaded from

http://oem.bmj.com/

Documents

Ventilation rate buildings building syndrome