Eur J Appl Physlol (1996) 73:452-458 © Springer-Verlag 1996

Urho M. Kujala • Simo Taimela • Tero Viljanen Helena Jutila • Jukka T. Viitasalo • Tapio Videman Michele C. Batti6

Physical loading and performance as predictors of back pain in healthy adults A 5-year prospective study

Accepted: 12 December 1995

Abstract We investigated muscle strength, aerobic power, and occupational and leisure-time physical loading as predictors of back pain in a 5-year follow-up study. A cohort of 456 adults aged 25, 35, 45 and 55 years, free of back pain, participated in measurements of anthropometric characteristics, aerobic power and muscle strength characteristics at baseline. The sub- jects' levels and types of physical activity and occupa- tional physical loading were also determined. At 5 years after the baseline examinations 356 of these subjects (78.1%) were reached by mail, and 262 of them (73.6%) properly completed and returned a question- naire including a detailed back pain history for the 5 years following the baseline measurements. Of this number 56 subjects (21%) who reported back pain ( > 30 on a scale from 0 to 100) and functional impair- ment during the 5-year follow-up composed the marked back pain group. Other subjects (n = 71, 27%) noting lesser symptoms were included in the mild back pain group; 135 subjects (52%) reported having had no back pain. The subjects with marked back pain were on average taller than the subjects without back pain, while no such difference was found in body

U.M. Kujala ( ~ ) . S. Taimela Unit for Sports and Exercise Medicine, Institute of Biomedicine, University of Helsmki, T6616 Sports Hall, Mannerheimintie 17, FIN-00250 Helslnki, Finland

T. Viljanen • H. Jutila Physical Fitness Research Unit of the Paavo Nurmi Centre, Uimahallinpolku 4, FIN-20320 Turku, Finland

J.T. Vitasalo Research Institute for Olympic Sports, UrJiversity Campus, FIN- 40700 Jyv~iskyl~, Finland

T. Vldeman Department of Health Sciences, University of Jyv/iskyla, FIN-40700 Jyv~iskyl~i, Finland

M.C. Battle Department of Physical Therapy, University of Alberta, Edmonton, Canada

mass. Heavy occupational musculoskeletal loading (P = 0.005) and high general occupational physical de- mands (P = 0.036) predicted future back pain. Leisure- time physical activity, aerobic power or muscle strength characteristics were not predictive of future back pain.

Key words Back pain- Occupational loading - Physical activity • Aerobic power • Muscle strength

Introduction

The effects of physical loading and factors of individual physical performance on back problems have been the focus of many studies, which have yielded conflicting results (Bigos and Batti~ 1990). It has been proposed that good muscle function can prevent back pain in some situations (Cady et al. 1979; Karvonen et al. 1980; Biering-Sorensen 1984), but such results have not been similar for subjects in other occupations or of both sexes (Biering-Sorensen 1984; Batti~ et al. 1989a, b).

There have been a relatively small number of studies which have collected baseline information on suspected risk factors for the purpose of identifying predictors of subsequent back symptoms (Gyntelberg 1974; Biering- Sorensen 1984; Leino et al. 1987; Troup et al. 1987). In these studies, in which the occurrence of back symp- toms was reported on questionnaire, or through inter- view, back symptoms have been found to be very common, the 1-year incidence ranging from 27% to 47% (Gyntelberg 1974; Troup et al. 1987). History of back symptoms was a strong indicator of future risk. Other factors associated with future complaints were measures of psychological distress and lower socio- economic status. Heavy occupational physical de- mands, greater age, smoking, chronic coughing, and leisure time inactivity, have been less markedly or less consistently associated with future back symptoms

(Gyntelberg 1974; Leino 1993; Karvonen et al. 1980; Videman et al. 1995).

A history of previous back pain is one of the most marked predictors of future back symptoms, and is correlated with other risk factors. The associations between back pain, physical performance character- istics, and external loading may involve complex inter- actions where primary causes and secondary effects are difficult to separate.

The aim of this study was to investigate the role of environmental physical loading and individual physical performance in healthy subjects without a history of recent or notable prior back problems. Spe- cifically, the objectives were to determine the 5-year incidence of back symptoms in previously healthy sub- jects from different age cohorts, and to investigate occupational loading, leisure time physical activity, an- thropometry, and physical performance characteristics as predictors of back symptoms over the subsequent 5-year period.

Methods

SubJects

A weighted (approximately 10% of each age group), random sample of 25, 35, 45, and 55-year-olds from Turku, Finland, was obtained from which to identify and select healthy subjects of differ- ent age groups, in which to study future back problems (Fig 1, Table 1). Of the 2007 people who were mailed a description of the study, a request for participation, and a baseline questionnaire, 1086 (54%) declined to volunteer for the physical examination portion of the study. Health reasons led to the exclusion of a further 465 (23%). This resulted in a complete baseline questionnaire and examination for 456 subjects who gave their reformed consent

453

prior to participation. At baseline, the subjects were invited to participate in a study in which physical fitness and health of the participants was investigated. The study was performed m accord- ance with the standard of ethics laid down in the Declaration of Helsmkl.

Health-related exclusion criteria included acute or chronic dis- ease, severe or recent back symptoms, or other musculosketetal symptoms that could interfere with the tests. At the laboratory the subjects were interviewed regarding their medical histories, and their

Population based and weighted random sample of Turku city population, representing ages 25, 35, 45 and 55, n=2265

4"~1 Not reached by mail, n = 258 4,

Reached by mail, n=2007

4"~1 Reported chronic diseases by mail, n=147 4"

Invited ~ participate in laboratory tests, n =1860

4'$1 Did not participate, n=1086 4,

Participated in laboratory tests, n =774 4, 4' $1 Reported chronic or acute diseases, previous severe 4' back pain, recent or current back pain, or symptoms thal 4' may interfere with the laboratory test results, n=318 4'

Complete baseline testing without any symptoms, n =456 4' 4'$1 Not reached at follow-up, n=100 4'

Reachedby mail at follow-up, n=356

4':~ Did not respond or responded too incompletely, n=94 4'

Properly responded at 5-year follow-up, n=262

Fig. 1 Process of recruiting subjects free of disease and back pain into the study

Table 1 SubJects of the follow- up study Age Sex

(years) (n) Subjects entering the laboratory for baseline measurements (n)

SubJects (%) free of SubJects reached at diseases and symptoms 5-year follow-up and m participating all responding properly the basehne measurements follow-up questionnaire 00

Men 353 191 (54%) 111 Women 421 265 (63%) 151 All 774 456 (59%) 262

25-55

25 Men 70 47 (67%) 24 Women 100 70 (70%) 35 All 170 117 (69%) 59

35 Men 115 71 (62%) 41 Women 146 98 (67%) 55 All 261 169 (65%) 96

45 Men 90 46 (51%) 23 Women 95 47 (49%) 29 All 185 93 (50%) 52

55 Men 78 27 (35%) 23 Women 80 50 (63%) 32 All 158 77 (49%) 55

454

electrocardiograms and resting blood pressures were obtained. As an additional safety precaution, a clinical examination was conduc- ted by a physician of all those aged 45 and 55 years prior to aerobic, strength, and performance tests, to ensure that there were no factors that may have placed these individuals at increased risk (Fig. 1, Table 1).

The subjects were questioned about previous and current back symptoms, and those who had reported previous severe back problems, musculoskeletal pain that had caused functional disability during the past year, or recent mild back pain that might have influenced the tests, were excluded. For example, sub- jects who had reported any back pain during the previous week or any musculoskeletal symptoms that had restricted their physical activity for more than 1 week during the past year, were excluded. Subjects with measurements that were influenced by pain were also excluded, so that the results obtained would be representative of healthy people. Back pain was the most common reason for exclusion from the different strength measure- ments, accounting for 28%-49% of exclusions in different age groups: it was followed by hypertension (diastolic blood pressure > 100 mmHg at rest) or a hypertensive reaction ( > 15 mmHg in-

crease of diastolic blood pressure, or increase of systolic blood pressure to over 240 mmHg during the submaximal exercise test preceding strength measurements), accounting for 13%-22% of exclusions, respectively. Exclusions at the laboratory were more common among the older ages, especially the men aged 55 years (Table 1).

Of the selected study group of 456 volunteers, 356 (78.1%) were reached by mail 5 years later, and 262 (73.6%) of them properly completed and returned the follow-up survey. The 262 subjects available for follow-up represented 57.5% of the original group of 456 participants, but only 12% of the base sample.

Baseline questionnaire

At baseline, the subjects completed a questionnaire mailed to them, which included questions on socao-economic status, health status, leisure-time physical activity, and occupational physical loading. On the basis of the responses, the following classifications were determined and used in analysis.

Soclo-economic status

This was classified based on the profession at baseline and socio- economic status was classified into the following main categories: executives, clericals, skilled workers, unskilled workers and farmers (Central Statistical Office 1972).

General occupational physical demands

The subjects themselves classified the general occupational demands for their work according to seven descriptions. For statistical ana- lyses these were combined into three categories: 1. Physically heavy or very heavy manual work 2. Light standing work or light to medium-heavy work involving movement 3. Not working or doing light sedentary work (for exact definitions see M~ilki~i 1983).

Occupational musculoskeletai loading

With no knowledge of other data, and based only on job titles, one of the authors (TV) who had special knowledge of the physical

demands of different occupations in Finland grouped the subjects into three categories with respect to work-related physical loading: 1 = mainly sedentary jobs (sitting at least two-thirds of working time) with minor, if any, lifting loads; 2 = mainly jobs with standing and walking including a variety of hght to moderate tasks with some bending, twisting, and lifting (loads less than 35 kg); 3 = heavy work including tasks with dally heavy lifting (loads more than 35 kg): and more frequent bending and twisting.

Leisure physical activity category

By combining the data on intensity and frequency, the following physical activity categories were determined: 1. Inactive or taking any activity less than once a week 2. Less intensive physical activity than in category 3, but exercising at least once a week 3. Physical activity that made the subjects sweat and become breathless at least once a week. A minimum mean duration of 25 mln of exercise per bout was required for inclusion in categories 2 and 3 (Kujala et al. 1994).

Mode of leisure physical activity

On the basis of the sports events reported (Kujala et al. 1994), the type of activity was classified into four categories 1. Activities reported (n = 34); 2. Walking only (n = 17); 3. Only typical aerobic training (n = t32) (the types of activity reported most often were: cycling 84, swimming 52, cross-country skiing 47, running/jogging 36 subjects) 4. Mixed training that included varied types of exercise with traumatizing manoeuvres (n = 79) (the types of activity reported most often were: volleyball 21, tennis 18, squash 11 subjects).

Baseline physical examination

The subjects anthropometrlc characteristics were described by body mass, stature, and body mass index (BMI, kilograms per metre square).

The vertical jumping height (height of rise of the body's centre of gravity), with counter movement and arm swing, was determined from the flight time of the three separate jumps. The flight time was determined by a contact mat (Ergo Jump: Bosco 1980). The recovery time between the attempts was 15 s. The best jumping height was used in further analyses. The coefficient of variation for duphcate measurements was found to be between 2.8% and 5.2% (Viitasalo et al. 1989).

The relative maximal isometric torque for trunk extension and flexion were measured with a dynamometer in a standing position (Viltasalo et al. 1977). After warming up, the subjects performed three maximal isometric trunk extensions and trunk flexions. Each trial took 3-5 s, with rest periods of 30 s between them. The mean of the best two results of the three trials was used for further analyses. For the calculation of the torque value, the distance between the trochanter major and force transducer (adjusted to the level of the inferior angle of the scapula) was measured and considered as the moment arm. The relative maximal torques were calculated by dividing the torque values by body mass (Newtons and metres per kilogram). The reproducibility of the tests of relative maximal Isometric strength of the trunk muscles expressed as the coefficient of variation between duplic- ated measurements was 2.0%-4.6% for the trunk extension and 2.6%-4.6% for the trunk flexion test (Viljanen et al. 1991).

The dynamic 30-s endurance tests of trunk extension and flexion were performed using individually-adJusted apparatus (M~lki/i 1983). In the trunk flexion test, the subjects kept their hands behind their necks while lying supine with the knees fixed in 90 ° flexion. The sit-ups were performed from a lying position to the point where the elbows touched the knees. The trunk extensions were measured on an examination table on which the subjects were lying prone with the trunk in a leaning position of 50 ° measured from the horizontal plane. The subjects were secured at their hips and ankles and they were free to move their unsupported upper body from a horizontal position (50 ° trunk flexion, the gravitational pull acting perpendicular to the upper body) to a full trunk extension (the gravitational pull acting at a 40 ° angle to the upper body). In both of these tests, the subjects performed as many repetitions as possible in 30s. MNki/i (1983) has reported a correlation coefficient of 0.92 between test-retest performed over a 1-year interval for the trunk flexion and 0.83 for the trunk extension test.

Maximal oxygen uptake

Maximal oxygen uptake (millilitres per minute per kilogram) was extrapolated indirectly on the basis of an electrically braked cycle ergometer test (pedal rate of 60 rpm) with three submaximal power outputs lasting 4 rnin each (see Andersen et al. 1971). The first submaxlmal power output used in the calculations should have produced a heart rate of approximately 120 beats . min -~, and the highest submaximal power output was approximately 85% of the age-adJusted maximal heart rate. The age-adjusted maximal heart rate was derived from the reference values of Sehger and Bartunek (1976). Maximal oxygen uptake was extrapolated using the maximal heart rate (Andersen et al. 197t: Seliger and Bartunek 1976).

Follow-up questionnaire

At the 5 year foliow-up, the subjects completed a questionnaire which included questions on SOClO-economic status, health status, leisure-tame physical activity, and occupational physical strain and also detailed questions on back pain. The subjects were asked for whole-life and last 5-year history of back pain (type, frequency, severity, and functional limitations). The severity of the worst pain episodes was rated on a scale from 0 to 100. One of the questions asked about the occurrence of back pain with pain radiating below the knee.

The functional impairment caused by the worst back pain episode was determined from the following questions, each scored on a 0 to 3 point scale. How much did you have difficulties when your back pain was at its worst during the past 5 years with (1) lifting or carrying a (grocery) bag, (2) bending or squatting, (3) walking for more than 1 km, (4) putting on socks and other clothes, (5) bedrest during the night, (6) standing up from a chair? Mild difficulties were given a score of 1, clear difficulties a score of 2 and inability to perform the activity received a score of 3. The maximal summary score was 18 (6 x 3).

The subjects who reported back pain during the 5-year follow-up period, and assessed that the intensity of their worst back pain was at least 30 on the scale from 0 to 100, and had a functional impair- ment score equal to or greater than 6 during the worst back pain episode were defined as the marked back pain group, Those subjects who did not report any back pain were defined as the no back pain group. The rest of the subjects were included into the mild back pain group.

The subjects were also asked a general question about whether they had clear difficulties or were unable to lift heavy objects during the worst back pain episode during the 5-year follow-up. If the

455

subjects had been prescribed time off work because of back pain before the follow-up period, this was also noted.

Statistical analyses

Statistical analyses included )~2 test or Fisher's exact test, analyses of variance, and two-way analyses of variance (sex. back pain) and covariance (age). Least-squares regression was used to produce age- and gender-adjusted averages of measurements done at the begin- ning of the study in the different back pain groups. Analyses were performed using version 6.04 of the Statistical Analysis System (SAS) software for microcomputers.

Results

To study whether our results would be affected by past back problems, we compared the subjects who had been prescribed rest because of back pain at some time earlier than 1 year before baseline and those who had not. Baseline measurements for the two groups were similar.

The incidence of back symptoms

During the 5 year follow-up 47% (n = 124) of the subjects reported having back pain, of whom 11% reported having back pain at least monthly. Back pain with pain radiating to the limb was reported by 17%, and 4 subjects (2%) had been hospitalized due to back pain during the follow-up.

Of the 47% of subjects (n = 124) reporting back pain during the 5-year follow-up, 82% rated the severity of their worst back pain to be at least 30 on the scale from 0 to 100, 42% gave their pain a rating of 60 or more, and 10% gave a rating of 90 or higher (Fig. 2). Of these 124 subjects reporting back pain, 47% reported at least a 6-point functional impairment during the worst back

120 - > 90 \ \ \ , ~ ,~ \ \ "

110" > ~ \ ~ , \ " bE

,~x,~x, 10-14 " ' " "\'" 90 - , ,-~ ,x\ '%) , \ e x \ #i ~ ~ ~ , ' ~ 4 , 6 0 - 9 0

80- 6-10 , ~ , ~ , ~ ~ " 70 ~ ~ ' ,

30-60 c~ 40 ~.,///~ (//,2

= 30 ~;~ ' /~2 0-6

o "/,//Y,/, Impairment Pain

Fig. 2 Functional impairment score and severity of back pain among the 124 subJects who reported back pain during the follow-up

456

pain episode, 26% reported at least a 10-point impair- ment and 6% a 14-point or greater impairment (Fig. 2).

During the 5-year follow-up, 21% of subjects experi- enced marked back pain, 27% had mild back pain, and 52% reported no back pain. Among the 56 subjects reporting marked back pain, all of whom rated the severity of their worst back pain on the scale from 0 to 100 to be at least 30, 37 subjects (66%) had a pain severity rating of at least 60, and 12 subjects (21%) had a rating of 90 or more. Among these subjects reporting marked back pain, 33 (59%) reported at least a 10- point impairment rating and 8 (14%) a rating of 14 points or more.

Indicators of future back symptoms

The 5-year incidence of marked back pain was 17% in the 25-year age group compared with 31% in the

45-year age group, but in a two-way ANOVA age was not statistically significantly associated with either the occurrence of back pain or back pain radiating to the leg during follow-up (Table 2). Back pain during the follow-up was associated with sex (P = 0.042, Z 2 test) such that there were proportionately more men in the mild back pain group and more women in the two other groups (no back pain or marked back pain; Table 2).

The subjects reporting marked back pain were on average taller than the subjects without back pain (mean 1.701 m vs 1.724 m, respectively, Table 3), but there were no significant differences in the body mass. There were no differences among the groups with no back pain, mild back pain and marked back pain, in any of the baseline aerobic, strength, or other perfor- mance tests (Table 3). In particular, we analysed differ- ent dynamic to isometric as well as trunk flexor to extensor strength test result ratios, but no association

Table 2 Back pain and referral pain by age and sex Age Sex

(years) (n) Back pain

None Mild Marked (%) (%) (%)

Referral pain to lower limbs (%)

25 Men 24 13 (54) 8 (33) 3 (13) 3 (13) Women 35 22 (63) 6 (17} 7 (20) 4 (11) All 59 35 (59) 14 (24) 10 (17) 7 (12)

35 Men 41 16 (39) 20 (49) 5 (12) 10 (24) Women 55 32 (58) 10 (18) 13 (24) 5 (9) All 96 48 (50) 30 (31) 18 (19) 15 (16)

45 Men 23 10 (43) 5 (22) 8 (35) 5 (22) Women 29 15 (51) 6 (21) 8 (28) 6 (21) All 52 25 (48) 11 (21) 16 (31) 11 (21)

55 Men 23 13 (57) 6 (26) 4 (17) 5 (9) Women 32 14 (44) 10 (31) 8 (25) 7 ~13) All 55 27 (49) 16 (29) 12 (22) 12 (22)

Men 111 52 (47) 39 (35) 20 (18) 23 (21) Women 15i 83 (55) 32 (21) 36 (24) 22 (15) All 262 135 (52) 71 (27) 56 (21) 45 (17)

25 55

Table 3 Age and sex adjusted means of different basehne anthropometric and physical capacity measurements in subjects with no subsequent back pain, with mild back pain, and with marked back pain

No Mild Marked back pain back pain back pain (n = 135) (n = 71) (n = 56)

F* P*

Body mass {kg) 69.8 Height (m) 1.701 Body-mass index (kg.m -2) 24.1 Predicted aerobic power (ml-mln-1.kg-1) 39.5 Jumping hmght (m) 0.228 Isometric trunk flexion torque (N.m.kg-1) 270 Isometric trunk extension torque (N.m.kg-1) 318 Dynamic trunk flexion (repetmons in 30 s) 14.8 Dynamic trunk extension (repetitions in 30 s) 20.2

71.7 69.9 1.29 0.28 1.708 1.724 3.25 0.040

24.5 23.5 2.23 0.11 38.7 40.8 1.18 0.31

0.227 0.225 0.16 0.85 263 268 0.33 0.72 317 315 004 0.96

14.8 14.3 0.42 0.66 20.8 20.0 1.10 0 33

* By 3-way ANOVA, results adjusted for age-group and sex (pain-group/age-group/sex)

between any of the ratios and future back pain was found. Nor was there an association between baseline leisure physical activity categories and future back pain.

Baseline general occupational physical demands did predict back pain (P = 0.036, Fig. 3). However, there was a clearer association between occupational mus- culoskeletal loading and future back pain (P = 0.005, Fig. 4). Socio-economic status did not predict future back pain.

All the Spearman correlation coefficients between height and general occupational physical demands and between height and occupational musculoskeletal loading for both the men and the women were low (r _< 0.2). Similarly, there were no significant correla- tions between any of the baseline aerobic, strength, or other performance test results and occupational phys- ical demands or between height and occupational mus- culoskeletal loading ( r < 0.15 for all correlations among both the men and the women).

457

Among the subset of 41 subjects (15.6%) having an occupation with '<heavy work including tasks with daily heavy lifting (loads more than 35 kg) and frequent bending and twisting" 44% reported having clear diffi- culties or were unable to lift heavy objects during the worst back pain episode during the 5-year follow-up. Among these 41 subjects we compared the 44% of the subjects with lifting difficulties to the 56% without marked lifting difficulties. After adjustments for age and sex the baseline aerobic or muscle performance characteristics were similar between the groups. The subjects with lifting difficulties were taller (mean ad- justed for age and sex, 1.693 m v s 1.732 m; F = 5.01, P = 0.031), than those without difficulties.

The future back pain with pain radiating to the lower limbs was associated with the baseline general occupa- tional physical demands (P = 0.005) and with the occu- pational musculoskeletal loading (P = 0.008), but not with the determinants of baseline leisure physical activ- ity or measurements of physical capacity.

80-

"~ 60-

2~

' .~ 40-

..13 20- Z

61% z.. (<,.4 .7/>>. I

/,, 7 ~A "//.."4

2" "/." ~,Z'..122% ~,,</.4.~ 18%

Mght

No Mild Marked

43%

~/ ) 32%

Y S X ~ ~ ' / /

" ~ ,

-/,/ Moderate Heavy

General occupational physical demands

Fig. 3 Baseline general occupational physical demands m three cat- egories and future back pain

62% 60-

Y, y/..,, 11-,

09 50- .. .

i l l l .

~ ' 40 .........

U) / X / V /

O 30-

i "/i I/ • ii ,i

20-

//,///

Z l o ~;'~ "//../,,,; , /, i

/ . / i /. i,~ 0 " , ' "

52%

LIg ht Moderate Heavy

Occupational musculoskeletal loading

Fig. 4 Basehne occupaUonal musculoskeletal loading in three categories and future back pain

Discussion

Our findings are in agreement with earlier studies re- porting an association between heavy physical work and back pain (Riihim~ki 1991). The association was particularly clear when the subjects' occupations were classified according to loading of the musculoskeletal system (Fig. 4). The subjects with occupations involv- ing a heavy physical loading were approximately 1.5 times more likely to report subsequent back pain than the subjects in occupations with light to moderate loading. Anthropometeric and physical performance measures failed to predict a risk of future back symp- toms in previously healthy, asymptomatic subjects. Height was the only exception, and although greater height was associated with subsequent back pain, parti- cularly marked pain, mean group differences were small (Table 3).

Our primary aim was to investigate predictors of back problems in previously healthy, asymptomatic adults. Because of the ubiquity of back symptoms and the likelihood of misclassification due to distant recall of pain history (Biering-Sorensen and Hilden 1984) we elected not to limit the study to subjects with no memory of ever having had previous back pain. However, all of the study subjects were symptom-free during the year preceding the baseline interviews and had no apparent functional limitations affecting performance testing. Selection bias due to low partici- pation rate at baseline with regards to back pain was unlikely since we did not tell the subjects that we were interested in back problems when inviting them to participate in our study at baseline. The participation rate to our 5-year follow-up of 262 of 356 subjects is satisfactory.

458

In line with our study, associations have been re- ported between greater height and hospitalizations for herniated lumbar intervertebral discs (Hrubec and Nashbold 1975), but findings have been less convincing with respect to more general back symptom complaints (Gyntelberg 1974; Biering-Sorensen 1984; Troup et al. 1987).

Similar to the findings of Leino et al. (1987), our study revealed no evidence to suggest that trunk muscle strength characteristics or strength ratios can be used to indicate risk of future back pain in healthy popula- tions. The finding that premorbid aerobic power was not a risk indicator is in agreement with an earlier study of predictors of general back pain complaints (Gyntelberg 1974), as well as prospective studies of work-related back pain complaints (Batti6 et al. 1989b; Ready et al. 1993). Level of participation in leisure time physical activities and types of sport involvement were not associated with future back pain either. Our study did not exclude the possibility, however, that physical training is a good means of maintaining an adequate functional level in back pain patients or to restore function in the rehabilitation of chronic low back pain.

In our study sample, due to statistical reasons, it was not acceptable to interpret multivariate models includ- ing all studied variables together. However, our cor- relation analyses showed that height, occupational physical loading and physical performance measure- ments did not correlate with one another, confirming that height and physical loading are independent pre- dictors for back pain and that the lack of association between baseline performance characteristics and fu- ture back pain is not due to differences in occupational loading. Our findings agree with previous studies that have reported that occupational physical activity does not increase the maximal oxygen uptake or muscle strength (Kujala et al. 1994).

In conclusion, with the exception of marginally greater height, anthropometric factors, aerobic power, strength measures, and other physical performance fac- tors were not associated with risk of future back pain in healthy adults. The study adds further evidence that heavy physical work increases the risk for back symp- toms. However preventive measures in heavy occupa- tions along would not resolve the back pain problem, because the majority of persons experiencing back pain work in lighter occupations.

Acknowledgement This study was financially supported by the Finnish Ministry of Education.

References

Andersen KL, Shephard R J, Denolm H, Varnauskas E, Masironi R (1971) Fundamentals of exercme testing. World Health Organ- lsation, Genova

BatU6 MC, Bigos S J, Fisher LD, Hansson TH, Jones ME, Wortley MD (1989a) Isometric lifting strength as a predictor of industrial back pain reports. Spine 14:851 856

Batti6 MC, BIgos S J, Fisher LD, Hansson TH, Nachemson AL, Spengler DM, Wortley MD, Zeh J (1989b) A prospective study of the role of cardiovascular risk factors and fitness in industrial back pain complaints. Spine 14:141 147

Biering-Sorensen F (1984) Physical measurements as risk indicators for low-back trouble over a one-year period. Spine 9:106 118

Biering-Sorensen F, Hilden J (1984) Reproducibility of the history of low-back trouble. Spine 9:280-286

Bigos S J, Batt16 MC (1990) Risk factors:industrial low back pain. In: Weinstein JN, Wiesel SW (eds) The lumbar spine. Saunders, Philadelphia, pp 846-858

Bosco C (1980) Sei und grande athleta. Vedlamo cosa dice l'Er- gojump. Pallavolo 5:34-36

Cady LD, BischoffDP, O'Connell ER, Thomas PC, Allan JH (1979) Strength and fitness and subsequent back injurms in firefighters. J Occup Med 21:269-272

Central Statistical Office (1972) Alphabetical list of occupations and classification of social class (in Finnish). Central Statistical Office, Helslnki

Gyntelberg E (1974) One year incidence of low back pain among male residents of Copenhagen aged 40-59. Dan Med Bull 21:30 36

Hrubec A, Nashbold BS Jr (1975) Epidemlology of lumbar disc lesions in the military in World War II. Am J Epidemiol 102:336 376

Karvonen MJ, Vaitasalo JT, Komi PV, Numml J, J~rvinen T (1980) Back and leg complaints in relation to muscle strength in young men. Scand J Rehabil Med 12:53 59

Kujala UM, Viljanen T, Taimela S, Vntasalo JT (1994) Physical activity, FO2 ..... and jumping height in an urban population. Med Sci Sports Exerc 26:889 895

Lelno P (1993) Does leisure time physical activity prevent low back disorders? A prospective study of metal industry employees. Spine 18:863-871

Leino P, Aro S, Hasan J (1987) Trunk muscle funcnon and low back disorders: a ten-year follow-up study. J Chron Dis 40:289296

M~ilkifi E (1983) Muscular performance as a determinant of physical ability in Finnish adult population. Social Insurance Institution, Turku

Ready AE, Boreskie SL, Law SA, Russell R (1993) Fitness and iifestyle parameters fail to predict back injuries in nurses. Can J Appl Phys 18:80-90

Rnhimiiki H (1991) Low-back pam, its orgin and risk indicators. Scand J Work Environ Health 17:81-90

Sehger V, Bartunek Z (1976) Internanonal Biological Programme results of investigations 1968 1974. Mean values of various indi- ces of physical fitness m the investigation of Czechoslovak popu- lation aged 12-55 years. CSTV Praha, Prague

Troup JDG, Foreman TK, Baxter CE, Brown D (1987) The percep- tion of back pain and the role of psychophysical tests of lifting capacity. Spine 12:645-657

Videman T, Sarna S, Batta6 MC, Koskinen S, Gill K, Paananen H, Gibbons L (1995) The long-term effects of physical loading and exercise lifestyles on back-related symptoms, disability, and spi- nal pathology among men. Spine 20:699-709

Viatasalo JT, Komi PV, Karvonen MJ (1977) Physical health and performance capacity and physical activity habits in conscripts in the beginmng of service at Air Force Communications School. Umversity of Jyv/iskyl~i, Jyv/iskyl~i

Viatasalo JT, Viljanen T, Kujala U (1989) Evaluation of vertical jumping tests. The XII International Congress of Biomechanics. 26-30 June, Los Angeles, USA. Available from the Department of Kinesiology, University of California at Los Angeles, Califor- nia, USA

Viljanen T, Viitasalo JT, Kujala UM (1991) Strength charactensncs of a healthy urban adult population. Eur J AppI Physaol 63: 43-47