where pij is the pressure on the sensor in row i, column j of the region, n is the number of sensors in the region registering nonzero pressure, and a is the area of an individual sensor (3.516 cm2). The Pliance mat is a matrix of square sensors of uniform dimensions, thus, the average pressure over a given region is equal to the average of the individual sensor pressures measured in that region.
Measures of foot and hand load.
Pressure on each Pedar insole measuring left foot and left hand pressure were converted to an applied load on each insole by multiplying each sensor pressure by the sensor area and summing over all sensors of the insole. Vector directions of the load calculated for the foot and hand could not be determined because the orientation of the plane of the pedal changes throughout the pedal revolution and the handlebar contact pressure was measured with the insole sensor wrapped around the cylindrical handlebar surface.
Spectral frequency analysis of the load on the left foot load time series revealed a clear spectral peak at a frequency that averaged 1.19 Hz, corresponding to a cadence of 71 rpm. Thus, participants did maintain a pedaling cadence close to the 70 rpm that was specified. Variation about this mean was small (SD of 2 rpm), and because the resistance of the ergometer was constant, absolute workload was considered to be constant for all participants. The periodicity of each foot load time series was well suited for a fit with a sinusoid function of the form:
where λ = [s/2πf], s is the sampling rate in hertz, and f is the frequency associated with the spectral peak of the time series (equivalent to the cadence). The frequency λ and amplitude α scaling of each sine function were determined in the spectral frequency analyses. The fit of the sinusoid functions to the measured foot load data were excellent as seen in the Figure 4 example.
It was assumed that the right and left foot load would be symmetric with a phase shift of one half of the pedal revolution. The foot load of the left and right foot combined was assumed to be equal to the sum of this sine function and the same sine function phase shifted by one half of a pedal revolution (see Fig. 4). Based on the assumption of left/right hand symmetry, average total hand load was assumed to be equal to two times the average left hand load measured on the Pedar insole.
The statistical models included independent variables for saddle, a covariate for body weight, and a body weight × saddle interaction. (The interaction between body weight and saddle was never statistically significant and was thus dropped from the final reduced models.) Probability of Type I error was set at 0.05 for all statistical tests unless denoted otherwise by a smaller P value. For dependent measures in which saddle reached statistical significance at the 0.05 level, Tukey’s Studentized range tests were conducted for comparisons between individual treatment means.
Pressure on the saddle.
Average pressure and peak localized pressure were calculated over two regions on all saddles. Measures of pressure over the full saddle were calculated from all sensors on the Pliance mat registering nonzero pressure. Measures of pressure on the perineal region were calculated in the same manner restricted to the region corresponding to the perineum as described previously. Peak and average pressure for both regions are shown by saddle in Figure 5.
The general linear modeling procedure was conducted for average and peak pressure over the full saddle and the perineal region of the saddle. The body weight × saddle interaction term was not significant in any model and was thus deleted in the reduced models reported. The covariate for body weight had a significant effect on the average full saddle pressure (F1,27 = 30.14, P < 0.001) and the peak and average pressure in the perineal region (F1,27 = 7.95, P < 0.01; F1,27 = 9.53, P < 0.01), but not on the peak full saddle pressure.
The traditional sport/racing saddle with the narrow protruding nose (saddle A) was associated with two times the average perineal pressure of the three nonprotruding nose saddles (37.2 kPa vs19.0, 19.4, and 16.4 kPa; F3,27 = 29.81, P < 0.001). The nonprotruding nose saddles (B, C, and D) were not significantly different with respect to their associated average perineal pressure. Peak perineal pressure was also significantly higher for the traditional saddle (70.4 kPa for saddle A vs 41.7, 41.5, and 42.0 kPa for saddles B, C, and D; F3,27 = 9.29, P < 0.001), but this relative difference between the traditional and nonprotruding nose saddles was not as large as that for average perineal pressure. Average pressure over the full saddle was significantly higher for saddle A than for saddle D (α = 0.05), whereas saddles B and C did not differ from A or D. Peak pressure over the full saddle was not different for saddles A and C or saddles B, C, and D. Groupings by statistical significance, from Tukey’s Studentized range test (α = 0.05), are shown in Figure 5.
Distribution of load on the pedals and handle-bars.
Average foot plus hand load as a function of saddle are shown in Figure 6. The body weight covariate had a significant effect on foot plus hand load (F1,27 = 8.91, P < 0.01). The variability in average foot load among participants was somewhat unexpected because pedaling resistance was constant for all participants. Body weight of the officers explained 47% of this variance (P < 0.01). Interestingly, body weight exhibited poor (and nonsignificant) correlation with average load measured at the hands (r2 = 0.06, P = 0.17). The hypothesized increase in relative loading on the feet and hands with the no-nose saddles was not supported. There were no significant differences among the saddles in the load calculated for the hands plus feet (F3,27 = 0.33, P = 0.81). Post hoc sample size calculations for this effect size showed that n = 88 participants per saddle group (N = 352) would be needed to demonstrate statistical significance at the 0.05 significance level with a power of 0.80.
These measures of hand and foot load are not measures of load in the vertical axis that functions to support the mass of the cyclist. This is particularly the case for the load on the hands, which may include gripping force around the handlebars. As a result, the absolute levels of the hand and foot load reported are less meaningful than the relative comparisons of these measures between saddles.
Bicycle saddle pressure is a function of numerous variables that cannot be comprehensively investigated in any single study. These variables are related to the anthropometrics and body position of the cyclist, the physical characteristics of the saddle, or the type of cycling. Physical characteristics of the saddle that influence the distribution of pressure on the cyclist are the geometry/shape and the cushion compliance. The traditional saddle evaluated in the present study (saddle A) was associated with an average pressure of 19.5 kPa over the full saddle and an average pressure over the perineal region of 37.2 kPa. Qualitatively, the traditional saddle examined in the present study was fairly stiff with relatively little cushion compliance compared to many saddles used by bicycle patrol officers. However, this traditional-design saddle was associated with an average full saddle pressure that was significantly higher than only one of the three nonprotruding nose saddles (saddle D). Thus, we conclude that the significantly higher pressure in the perineal region associated with the traditional saddle is more likely influenced by the saddle geometry and shape than by its cushion properties.
The anthropometry of each cyclist’s perineal region could not be individually characterized. This necessitated a method for inferring the spatial location of the perineum with respect to the saddle so that the spatial distribution of pressure registering on the pressure mat could be mapped to the perineum. The method adopted to identify this spatial location was based on landmarks assumed to correspond with the ischial tuberosities. The representations of average saddle pressure shown in Figure 2 are two-dimensional planar representations of pressure registered on bicycle saddles that have nonplanar surfaces. As a result, some portion of the pressure registered on the sensors of the mat assumed to be in the perineal region of the cyclist may not have created harmful pressure on the pudendal nerves and vasculature. Portions of the inner and anterior thigh make contact with the saddle in concert with the pedaling dynamics, and we were not able to separate the inner/posterior thigh contact pressure from that which was truly pressure in the perineum. This phenomenon is even more likely with the nonprotruding nose saddles that are “rounded-off” at the front where the saddle nose would be present on a traditional saddle. Thus, we believe that the true perineal pressure reported in this study may be over-represented for all saddles, but this over-representation is likely greater with the nonprotruding nose saddles (saddles, B, C, and D) where the inner thigh comes into contact with the rounded front portion of the saddle.
The bicycle ergometer adjustments of saddle height and handlebar position were selected by the individual participants based on their own subjective preference of cycling position. In the evaluation of saddle designs other investigators (Bressel and Larson; 3) have standardized saddle height as a fixed percentage of inseam height. However, the commonly accepted recommendations for optimal saddle height/inseam height ratio have been based on cycling performance using traditional shaped saddles with a protruding nose design. The validity of these recommendations when applied to other radically different saddle designs (such as saddles B, C, and D in the present study) is unknown. For this reason, we chose a method of saddle height selection based on the cyclists’ subjective preference and comfort over standardizing saddle height based on the cyclist’s anthropometry. Other than body weight, no other anthropometrics were recorded; thus, it is not known whether participants’ self-selected saddle heights within saddle group were consistent with respect to anthropometrics such as inseam height.
The design of two of the nontraditional saddles introduced the phenomenon of hammocking, which has been described by Ferguson-Pell and Cardi (6). Hammocking occurs when the pressure sensor mat spans or hammocks over a split, cutout, or discontinuous region of the saddle cushion and the sensor registers pressure on the mat over this region where no cushion material is present underneath. Saddles C and D both had regions of discontinuity of the cushion surface along the midline of the saddle. Saddle C exhibited a region of slight hammocking that is evident in Figure 2c in the region spanning the midline of the saddle where the two cushioned halves of the saddle are split. This hammocking is a measurement artifact relevant to saddles C and D, but not saddle B or the traditional saddle, neither of which had discontinuities in their cushion surface. The hammocking artifact also suggests that the pressures measured in association with the nonprotruding nose saddles (C and D) were overestimated relative to the pressure associated with the traditional saddle, which still exhibited two times greater pressure in the perineal region.
Moes (14) noted that the pressure threshold for decubitus is as low as 7 kPa for long-term pressure on the skin. The average levels of perineal pressure measured in the present study were substantially higher than 7 kPa. The traditional saddle was associated with an average perineal pressure between 34 and 41 kPa. The saddles without protruding noses were associated with an average perineal pressure of approximately 18 kPa. Data presented by Armstrong (2) indicates an exposure duration tolerance of approximately 150 min at 34 kPa. This explains the development of the “saddle sores” often experienced by cyclists after several hours in the saddle.
The present findings in regard to the relationship between degree of protrusion of the saddle nose and perineal pressure are in agreement with studies of Jeong et al. (11) and Schwarzer et al. (18). The latter of these studies revealed an 82.4% reduction in transcutaneous penile oxygen pressure with a traditional racing saddle design and only a 20.3% reduction with a wide saddle without a protruding nose. Jeong et al. (11) revealed a substantially greater decrease in penile blood flow associated with sitting on a saddle with a long narrow nose than with a wide saddle with a lesser protruding nose. These studies and the present study implicate the shape and protrusion of the saddle nose and the resulting distribution of pressure in the perineum in the compression of the vasculature supplying the penis.
The present data do not support the hypothesis that the absence of a saddle nose causes a shift in the distribution of load among the saddle, pedals, and handlebars. However, it is well recognized that the present study considered stationary cycling in which bicycle stability, handling, and maneuverability were not relevant. Anecdotal reports from police patrol officers and recreational cyclists suggest that the absence of a protruding nose may adversely affect bicycle handling and maneuverability. Future studies should be based on road cycling to consider handling and maneuverability issues in the evaluation of the effectiveness of saddle designs. A longitudinal intervention study is also needed to demonstrate the benefit of saddles without a protruding nose.
We conclude that bicycle saddle designs without a narrow protruding nose significantly reduced pressure distributed in the perineal region of the cyclist during stationary cycling. Based on previous work that indicated a relationship between pressure on the saddle nose and the quality of nocturnal erectile tumescence (16), this reduction in perineal pressure is believed to reduce the risk of erectile problems associated with occupational cycling.
Future work will examine the benefits of saddles without narrow protruding noses in a prospective longitudinal study design in which the cyclists use these saddles in actual road cycling over a longer study period. The longitudinal study design is critical in demonstrating the health benefits of saddles without protruding noses and in determining cyclists’ acceptance of these saddle designs. Evaluations of the effect of nonprotruding nose saddle designs on bicycle maneuverability, handling, stability, and weight distribution should be conducted in actual road cycling.
The authors wish to acknowledge the assistance of Roger Baker for fabricating the seat post support modification to the ergometer, Peter Shaw for statistical support, and Dan Habes, Greg Cutlip, and Chris Pan for their commentary on the manuscript. Special gratitude is also extended to the International Police Mountain Bike Association for their receptiveness to our study of this topic.
1. Andersen, K. V., and G. Bovin. Impotence
and nerve entrapment in long distance amateur cyclists. Acta Neurol. Scand.
2. Armstrong, T. J. Mechanical considerations of skin in work. Am. J. Ind. Med.
3. Bressel, E., and B. J. Larson. Bicycle seat designs and their effect on pelvic angle, trunk angle, and comfort. Med. Sci. Sports Exerc.
4. Desai, K. M., and J. C. Gingell. Hazards of long distance cycling. Br. J. Med.
5. Dickson, T. B. Preventing overuse cycling injuries. Physician Sportsmed.
6. Ferguson-Pell, M., and M. D. Cardi. Prototype development and comparative evaluation of wheelchair pressure mapping systems. Assistive Technol.
7. Goldstein, I. Editorial commentary. J. Androl.
8. Goodson, J. D. Pudendal neuritis from biking. N. Engl. J. Med.
9. Groenendijk, M. C., H. C. C. M. Christiaans, and C. M. J. van Hulten. Sitting comfort on bicycles. In:Contemporary Ergonomics
, E. J. Lovesey (Ed.). Proceedings of the Ergonomic Society’s 1992 Annual Conference, Washington, DC: Taylor and Francis, 1992, pp. 551–557.
10. Hershfield, N. B. Pedaller’s penis. Can. Med. Assoc. J.
11. Jeong, S. J., K. Park, J. D. Moon, and S. B. Ryu. Bicycle saddle shape affects penile blood flow. Int. J. Impot. Res.
12. Marceau, L., K. Kleinman, I. Goldstein, and J. Mckinlay. Does bicycling
contribute to the risk of erectile dysfunction
? Results from the Massachusetts Male Aging Study (MMAS). Int. J. Impot. Res.
13. Mellion, M. B. Common cycling injuries management and prevention. Sports Med.
14. Moes, C. C. M. The development of a pressure distribution measuring device for various person-product contact areas. In:Contemporary Ergonomics
, E. D. Megaw (Ed.). Proceedings of the Ergonomic Society’s 1989 Annual Conference. New York: Taylor and Francis, 1989, pp. 349–354.
15. Oberpenning, F., S. Roth, D. B. Leusmann, H. van Ahlen, and L. Hertle. The Alcock syndrome: temporary penile insensitivity due to compression of the pudendal nerve within the Alcock canal. J. Urol.
16. Schrader, S. M., M. J. Breitenstein, J. C. Clark, B. D. Lowe, and T. W. Turner. Nocturnal penile tumescence and rigidity testing in bicycling
patrol officers. J. Androl.
17. Schwarzer, U., W. Wiegand, A. Bin-Saleh, et al. Genital numbness and impotence
rate in long distance cyclists. J. Urol.
18. Schwarzer, U., F. Sommer, T. Klotz, C. Cremer, and U. Engelmann. Cycling and penile oxygen pressure: the type of saddle matters. Eur. Urol.
19. Silbert, P. L., J. W. Dunne, R. H. Edis, and E. G. Stewart-Wynne. Bicycling
induced pudendal nerve pressure neuropathy. Clin. Exp. Neurol.
20. Solomon, S., and K. G. Kappa. Impotence
: a seldom-reported connection. Postgrad. Med.
21. Sommer, F., D. Konig, C. Graft, et al. Impotence
and genital numbness in cyclists. Int. J. Sports Med.
22. Weiss, B. D. Clinical syndromes associated with bicycle seats. Clin. Sports Med.
Keywords:©2004The American College of Sports Medicine
BICYCLING; ERECTILE DYSFUNCTION; IMPOTENCE; GROIN PRESSURE