See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/40766295

Skyscraper running: Physiological and biomechanical profile of a novel sport

activity

Article in Scandinavian Journal of Medicine and Science in Sports · April 2011

DOI: 10.1111/j.1600-0838.2009.01043.x · Source: PubMed

CITATIONS READS

25 439

5 authors, including:

Alberto E Minetti Dario Cazzola

University of Milan University of Bath
162 PUBLICATIONS 7,853 CITATIONS 74 PUBLICATIONS 609 CITATIONS

SEE PROFILE SEE PROFILE

Elena Seminati Giulio Sergio Roi

University of Bath University of Bologna
38 PUBLICATIONS 483 CITATIONS 104 PUBLICATIONS 3,432 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Elena Seminati on 01 May 2018.

The user has requested enhancement of the downloaded file.

Scand J Med Sci Sports 2009 & 2009 John Wiley & Sons A/S
doi: 10.1111/j.1600-0838.2009.01043.x

Skyscraper running: physiological and biomechanical profile of a

novel sport activity
A. E. Minetti1, D. Cazzola1, E. Seminati1, M. Giacometti2, G. S. Roi3
1
Department of Human Physiology, Faculty of Medicine, University of Milan, Milan, Italy, 2Federation for Sport at Altitude, Biella,
Italy, 3Education and Research Department Isokinetic, Bologna, Italy
Corresponding author: Alberto E. Minetti, Department of Human Physiology, Faculty of Medicine, University of Milan, Via
Mangiagalli 32, 20133 Milano, Italy. Tel: 139 02 5031 5427, Fax: 139 02 5031 5430, E-mail: [email protected]
Accepted for publication 20 August 2009

Skyscraper running is here analyzed in terms of mechanical resources particularly needed at the beginning of the race. The
and metabolic requirements, both at the general and at the total mechanical power was found to be partitioned among:
individual level. Skyscraper runners’ metabolic profile has the
. fraction devolved to raise the body center of mass
been inferred from the total mechanical power estimated in WSTA:EXT ¼ 80:4  2:9%,. the need to accelerate the limbs
36 world records (48–421 m tall buildings), ranked by with respect to the body WSTA:INT ¼ .4:5  2:1%, and run-
gender and age range. Individual athlete’s performance ning in turns between flights of stairs WTUR ¼ 15:1  2:0%.
(n 5 13) has been experimentally investigated during the At the individual level, experiments revealed that these
Pirelli Vertical Sprint, with data loggers for altitude and athletes show a metabolic profile similar to middle-distance
heart rate (HR). At a general level, a non-linear regression runners. Furthermore, best skyscraper runners maintain a
of Wilkie’s model relating maximal mechanical power to constant vertical speed and HR throughout the race, while
event duration revealed the gender and age differences in others suddenly decelerate, negatively affecting the race
terms of maximum aerobic power and anaerobic energy performance.

Running uphill on steep emergency stairs, run-up energy rate) has to be strictly proportional to the
races as they are usually called today, is a rapidly total mechanical work (WTOT or power) generated
expanding sport performed on the tallest buildings of by muscles during the ascent, the last being an easy
the planet. Running on stairs has been an interesting variable to calculate.
motor activity since Rodolfo Margaria’s time, when Another attractive aspect relates to the presence, in
he designed the rapid ascent test to evaluate the most skyscrapers, of handrails that maximize the
individual maximum anaerobic power (Margaria muscle mass involved and, consequently, the me-
et al., 1966). Only a few studies have previously dis- chanical/metabolic power of the ascent, conferring
cussed stair climbing, some of them considering the race with a feel of a global, maximal effort as in
kinetics and kinematics (Mc Fadyen & Winter, rowing. Because the duration of the events ranges
1988; Yu et al., 1997; Larsen et al., 2009) or meta- from a few dozens of seconds to 14 min and runners
bolic aspects of slow walking on stairs, or under attend many different races, both anaerobic and
particular conditions such as running while wearing aerobic skills are simultaneously required. Thus, the
firemen robes (O’Connel et al., 1986; Teh & Aziz, athlete’s choice in terms of the sustainable ‘‘engine
2002). The particular appeal in this new sport dis- set-point’’ is crucial to the overall performance, as an
cipline resides in the fact that mostly positive work is excessive initial power could negatively affect aerobic
carried out (Minetti et al., 1994), that the elastic pathway enzymes and jeopardize the rest of the
energy storage and the consequent release in running competition.
are practically nil at steep gradients, and that the Run-up races are organized on buildings of very
work to increase the gravitational potential energy different heights, and they allow to test predictions
[the prevalent portion of the external mechanical about the maximum mechanical power sustainable
work (WEXT), needed to accelerate and lift the for a given exercise duration (e.g. Wilkie, 1980) in a
body center of mass (BCOM)] considerably exceeds wide range of performances. This analysis will also
the work to move limbs with respect to the center of provide the ‘‘typical’’ profile of run-up male and
mass (the mechanical internal work, WINT). Thus, female athletes of very different ages in terms of the
the required metabolic energy (or the metabolic aerobic and anaerobic resources available.

1
Minetti et al.
The aims of this paper are (1) to define a research i.e. the .sum of the power necessary to run up the
methodology for this new sport, (2) to measure the stairs (WSTA ) and the one related to .running in turns
physio-mechanical variables of a group of athletes between successive flight of stairs (WTUR ).
during a real run-up race and (3) to infer from them The first term is classically partitioned into the
the climbing strategy, if any. We will introduce the external and the internal portion of the mechanical
topic, in the following, by reviewing and analyzing power
world records in this expanding sport activity. . . .
WSTA ¼ WSTA:EXT þ WSTA:INT ½2
.
WSTA:EXT is estimated as
Analysis of world records . m gDh
WSTA:EXT ¼ ½3
As shown in Fig. 1, races have been organized so far Dt
in buildings of very different height, ranging from a where m, g, Dh and Dt are the subject mass (made
few dozen meters to the tallest skyscrapers on the equal to 70 kg), the gravity acceleration, the height of
planet (KL Tower, Malaysia, 421 m). The event the race inside the building (m) and the race time (s),
duration ranges from 50 s to about 14 min and the respectively. We disregarded both the vertical and the
race conditions vary considerably, not just for cli- forward kinetic energy changes of the BCOM because
matic reasons. In some cases, a short, level approach they are assumed not to affect the overall mechanical
to stairs is included and the staircase steps, while work, being ‘‘buried’’ in the monotonically ascending
being quite similar, are not exactly of the same size/ curve of the total energy of BCOM when running at
geometry in the different buildings. Runners are very steep gradients (Minetti et al., 1994).
often divided into groups, to avoid overcrowding .
The term WSTA:INT reflects the mechanical internal
the race terrain, and the group size can affect the power necessary to accelerate limbs with respect to
overall performance. In addition, the stairs’ width BCOM (Cavagna & Kaneko, 1977). Normally obtained
affects the usage of handrails as pushing aids, and the by processing kinematics data, here, we estimate its real
number of floors implies a different number of value using a model equation that has been tested
‘‘running turns’’ for the same vertical distance tra- previously for gradient locomotion (Minetti, 1998)
velled.  2 !
As anticipated in the introduction, run-up races . .2 d
WSTA:INT ¼ m f s 1 þ q ½4
appear to be more easy to analyze mechanically than 1d
other sports as, for example, level running, because .
the power to move vertically is expected to be the where f is the stride frequency (Hz), s is the (diagonal)
predominant fraction of the total power needed. In speed (m/s) on the stairs, d is the duty factor, i.e. the
order to consider the most comprehensive list of fraction of the stride period at which one foot is in
determinants, though, we modelled the total mechan- contact with the ground and q reflects the inertia
ical power required to complete the ascent as properties of the four body limbs. Depending on the
. . . building height, the race time, the reported number of
WTOT ¼ WSTA þ WTUR ½1 steps and previous data from our group on gradient

1000 500

900 450

800 400
total mechanical power (W)

700 350
race height (m)

600 300

500 250

400 200

300 150

200 100

100 50

0 0
0 100 200 300 400 500 600 700 800 900
race duration (s)
.
Fig. 1. Current world records of male athletes in run-up races are represented as the ‘‘minimum’’ mechanical power (WTOT )
needed to set them (solid circles, see text), together with the building height (minus symbols) shown for each record on . the
right-hand side ordinate, as a function of race duration. The curve represents a non-linear regression of world records (WTOT )
based on Wilkie’s model (see eqn. [5]).

2
 Skyscraper running physio-mechanics
running (Minetti et al., 1994; Minetti, 1998), we esti- 100%
.
mated for each record f and s, while d and q were
assumed to be equal to 0.45 90%
. (extrapolated for a gradient
of about 50% from the WINT model) and 0.15, respec-
80%
tively. Stride frequency was deducted from the number

work or power fraction (%)

of stair steps and from observing that athletes run on 70%
.
every other step, and s was estimated by dividing the
vertical speed by sin(a tan(i)), where i is the stairs’ 60%
gradient (assumed to be equal to the one we measured vertical
on the Pirelli building, about 50%). 50% turning
The mechanical power involved in running in turns internal
. 40%
between flights of stairs (WTUR ) and its metabolic
energy consumption have never been studied in the 30%
past. A set of preliminary experiments
. provided the
metabolic equivalent of WTUR (see Appendix). A 20%
curvature radius of 1 m describing the body trajec-
tory to the next floor, as measured on the Pirelli 10%
building in Milan, has been assumed for all the
buildings involved in world run-up records. Because 0%
0 100 200 300 400 500
of this methodological
. approach, we can expect that
race height (m)
estimated WTUR reflects the sum of the external and
the internal power of running in circles. . Fig. 2. The three components
. of the estimated
. mechanical
The total mechanical power (WTOT ) needed to power
. (vertical: WSTA:EXT , turning: WTUR . , internal:
achieve the world records on 36 skyscrapers, calcu- WSTA:INT , see text) are shown as fractions of WTOT , for the
different race heights.
lated according to the above equations, is shown as
solid circles in Fig. 1. The prediction is supposed to
be reliable due to the expected absence of the elastic subjects reach exhaustion at the end of the exercise
energy stored and released during the contact phase  t
. B A  t 1  et
(Minetti
. et al., 1994). In any case, the estimated WMECH ¼ A þ  ½5
WTOT and the related metabolic energy consumption, t t
obtained by dividing by the muscle contraction where A is the maximum long-term mechanical work
efficiency, represent the ‘‘minimum’’ work rate and rate (W), B is the mechanical equivalent of the
energy consumption . for each ascent. Additional available energy from anaerobic sources (J) and t is
components of VO2 could include, for example, the the time constant (s) reflecting the inertia of the
effect of antagonist muscles, the need to stabilize the system. Wilkie modelled this equation to be accurate
trunk during uphill running and the acceleration at for durations ranging from 40 s to 10 min, which very
the beginning of each flight of stairs. . closely match run-up competitions.
When the three main components of WTOT are Figure 1 also shows a curve obtained by perform-
plotted vs race height, as in Fig. 2, it appears that their ing a non-linear
. regression based on Wilkie’s model
contribution is quite constant, being . 80.4 . 2.9%, on the WTOT data. The software, written in LabView
4.5  .2.1% and 15.1  2.0% for WSTA:EXT , WSTA:INT programming language (downloadable at albertomi-
and WTUR , respectively. netti.it/MathModCompMeth), has been tested with
While in the early days run-up participants were mock data series created according to known A, B
amateur runners, more recent records . have been and t values. The resulting coefficients were
established by professionals; thus, WTOT data in Fig. A 5 416.4 W, B 5 23002.3 J and t 5 5.5 s.
1 represent performances of top male athletes in the A and B, once converted into metabolic units
age range 20–35 years who often attend multiple events according to an efficiency value for muscle contrac-
on skyscrapers with very different heights (and race tion of 0.25, define the metabolic profile of this class
durations). It is apparent that the total mechanical of athletes. These parameters correspond to the
power generated is much greater for smaller buildings, maximum sustainable aerobic power above resting
where the shorter performance allows to exploit more [1.67 metabolic kW or 71.4 ml O2/(kg min)] for an
powerful, high-energy phosphate sources. assumed body mass of 67 kg, in agreement with
Wilkie (1980) attempted to capture this phenom- typical values of middle-distance runners (Saltin &
enon by proposing an equation predicting . the avail- Astrand, 1967) and the anaerobic capacity (92 meta-
able average mechanical
. power ( W MECH (W) – the bolic kJ, equivalent to 66 ml O2/kg, a reasonable
same as our WTOT ) as a function of the event dura- value indeed for elite athletes), respectively. As men-
tion [t (s), same as our Dt] with the assumption that tioned by Wilkie (1980), t is much shorter than the

3
Minetti et al.

Anaerobic Sources B (J) Maximum Aerobic Mechanical

500

n=34
400 n=29
n=30 n=28
Power A (W) n=27
n=26
300
n=23 n=31 n=28 n=29 n=14
200 n=20
n=14
100

0
25000
Mechanical Energy from

20000

15000

10000

5000

0
12

10

8
Tau (s)

0
0 10 20 30 40 50 60 70 80
Age (yrs)

Fig. 3. Effects of age and gender (n is the sample size) on the

mechanical equivalent of the maximum aerobic work rate, of
the anaerobic work capacity and time constant (A, B and t,
respectively, in eqn. [5]), as obtained using non-linear regres-
sions based on Wilkie’s model. The B and t values in the
oldest age ranges have not been shown because the paucity
of records in smaller skyscrapers for those athlete cohorts
suggests caution about their reliability.
Fig. 4. Upper panel: age distribution of participants of the
Empire State Building Run-Up Race held in 2004. Middle
panel: vertical mechanical power of female and male winners
time constant of the metabolic transient (typically in of run-up races held in four subsequent years. Lower panel:
the region of 40 s) because it reflects the speed at vertical mechanical work per heart beat of male winner of the
which the mechanical output increases at the begin- four run-up races (see the text for details).
ning of the exercise.
The analysis of world records also gave us the Building Run-Up in New York (320 m height, 1576
opportunity to indirectly study the metabolic profile steps, 86 floors), which has been running since 1978.
of athletes of different genders and ages. The same Because of the relevant number of male and female
non-linear regression applied above on the absolute athletes of all ages attracted by that race (upper panel
winners was used on clusters of data pertaining to of Fig. 4), the variability in terms. of event records
best male and female athletes (female mass assumed from 2003 to 2006 (expressed as WSTA:EXT and shown
to be 55 kg) within different age ranges categories in the middle panel, Fig. 4) is further reduced. Again,
(from 10 to 80 years, step 10 years). The results are the effect of age is apparent in reducing the available
shown in Fig. 3, where the decay of A and B with mechanical power. Interestingly, even the oldest
increasing age is apparent. participant (aged 91 in 2003) showed worse perfor-
The data scatter observed in Fig. 1 is caused by the mances in successive competitions (the three right-
heterogeneity of building heights and race features. most points for males).
The variability can be reduced by considering the Inferences about the amount of mechanical work
results from just one race, performed in a given time sustained by each heart beat (J/beat) as a function of
span. The oldest competition was the Empire State age have been determined by dividing the mechanical

4
 Skyscraper running physio-mechanics
.
vertical power (WSTA:EXT ) by the heart rate (HR) 305, Garmin (Garmin International Inc., Olathe, Kansas,
(80% of the maximum, estimated according to age, USA)) at a lower resolution (1 m) and sampling rate (1 Hz
minus the basal value, assumed to be equal to maximum). Also, HR was monitored during the race (Edge
305 by Garmin; Vantage, RS800 and S810 by Polar, Vantage
55 b.p.m.). The lower panel of Fig. 4 shows that up & Polar, Oulu, Finland). Ascension and HR data of other four
to the age of 65–68 years in males, the work per beat elite athletes [including the winner (Thomas Dold) of this and
is quite constant, suggesting that the HR could be the many other run-ups] were provided from their own monitor-
most crucial factor in mechanical power reduction ing equipment after the race.
with age. The consistent decline for older athletes Lactate was measured (Lactate Analyser YSI Sport, Yellow
Spring, Ohio, USA) 3–6 min after arrival in 21 male amateur
would imply that other factors, such as the decreases athletes (mass 68.2  10.0 kg, stature 173.8  6.6 cm, age
of both the maximum HR and the stroke volume or a from 26.5 to 68.6 years) to check their anaerobic status and
lower oxygen extraction, act in combination to to estimate the lactate contribution to the work production.
reduce the available mechanical power. In order to mathematically describe the two-segment shape
of the time course of vertical speed observed in many athletes
(see ‘‘Results’’), a statistical algorithm fitting a single data set
with two successive regression lines was used (Jones &
Experiments during the Pirelli run-up Molitoris, 1984). The method searches, among all possible
two-line combinations, the optimum one in terms of the
As shown in the preceding paragraphs, exercise minimum residuals across the entire data set and statistically
physiology and locomotion biomechanics provide assesses whether two lines fit better than a single one. A result
sufficient information to analyze run-up performance that is derived is the optimum breakpoint, i.e. the point in the
just from record times. These inferences are valid for abscissa (and the corresponding ordinate) at which the second
the average runner, be it a male or a female of a given regression line comes into action. The method has been
translated into a computer program (LabView, National
age. Individual athletes, though, need to be mon- Instrument, Austin, Texas, USA) and additional features
itored during training and sport events by a proper have been added. Because it is expected that points lying on
research protocol. In the following, we describe a a (concave or convex) monotonous curve are better fitted by
simple set of preliminary measurements based on the two lines than by a single one, we included a subroutine
above analysis of run-ups that we used during a real checking to observe whether a second-degree polynomial fits
better than two adjacent lines, again based on the minimiza-
race. tion of the sum of residuals (a more detailed description and
the program is downloadable at albertominetti.it/MathMod-
CompMeth). The single line and the second-degree polynomial
Methods represent continuous time courses of the vertical speed, while the
two-segment model describes a discontinuous performance.
The ascent speed of 13 male athletes (see anthropometric data While a parabola regression provides three parameters, the
in Table 1), who gave their written informed consent for the two-segment model would involve four of them (two intercepts
experimental procedure, was investigated during the run-up and two slopes) but, due to the constraint imposed by the
races on the Pirelli building (121 m height, 710 steps, 30 floors) intersection, the effective parameters reduce to three, and the
in Milan on February 24 2008 and on March 1 2009. The adherence of the two models to the same data can be compared.
study was approved by the ethics committee of the University
of Milan. Altitude was measured on five athletes using an
altimeter1logger device, designed to monitor model aircrafts
and rockets, capable of a 0.4 m resolution and a sampling rate Results
of 10 Hz (LoLo/Alti2, Roman Vojtech, http://www.lomcov-
ak.cz). The ascent of other four athletes was measured by the In the male amateur athletes, the average lactate con-
internal barometric altimeters of GPSs (Geko 301 and Edge centration at the end of the race was 5.3  1.4 mM

Table 1. Antropometric data of the investigated run-up runners are shown in the table, together with the estimated partitioning of their total mechanical
power (see text)
. . . .
Subjects Age Body Stature Ascent Vertical WSTA:EXT WSTA:INT WTUR WTOT
(years) mass (kg) (m) time (s) speed (m/s) (WMECH) (WMECH) (WMECH) (WMECH)

1 23 69 1.78 188 0.642 435.7 30.663 75.26 541.6

2 22 68 1.79 209 0.579 386.9 22.121 66.85 475.9
3 38 68 1.77 222 0.544 363.6 18.352 62.81 444.8
4 31 60 1.65 255 0.473 279.0 10.647 48.19 337.8
5 28 59 1.64 270 0.447 259.6 8.872 44.85 313.3
6 37 72 1.76 249 0.484 342.8 13.721 59.22 415.8
7 42 77 1.78 258 0.468 354.3 13.239 61.20 428.7
8 28 72 1.52 314 0.384 272.2 6.867 47.02 326.1
9 48 75 1.76 329 0.367 270.6 6.219 46.75 323.6
10 64 69 1.75 335 0.360 244.5 5.419 42.24 292.1
11 58 78 1.67 362 0.333 255.6 4.847 44.16 304.6
12 47 72 1.74 365 0.331 234.1 4.372 40.45 279.0
13 49 53 1.68 428 0.282 147.0 1.997 25.40 174.4

5
Minetti et al.

Fig. 5. Two examples of ascent profile and concurrent heart rate time course (upper and lower panels, respectively) are shown
for two typical athletes: a discontinuous profile and a very steady ascent (left- and right-hand side, respectively). It can be noted
that the discontinuous profile corresponds to a heart rate overshoot that could jeopardize the rest of the race performance.

(range 2.50–7.75), with no significant effect of race

speed on it.
Table 1 reports the partitioning of the total me-
chanical work rate for the 13 individually studied
runners according to the rationale described in the
introduction (‘‘Analysis of world records’’). Nine of
them, monitored through digital altimetry during the
ascent, showed a sudden change of vertical speed, as
significantly detected by the two-segment regression.
For eight of them, the change was a reduction of
vertical velocity (see an example in Fig. 5, column A),
while one subject increased his vertical velocity in the
last part of the race. This sudden change was con-
firmed by the ratio between the residuals related to a
second-degree polynomial and the residuals obtained
with the two-segment model, which was found to be
greater than one in nine cases. Two of these nine
athletes showed high ratio values, underlining their
considerable speed reduction, as shown by the corre- Fig. 6. The ascent profile factor is plotted against the
lation coefficient of the two-segment model (Fig. 6). average vertical speed for seven under-40 (solid squares)
and six over-40 (open squares)-year-old athletes. The ascent
In the remaining four athletes, who were the highest profile factor has been calculated as the ratio between the
in the race ranks for both the under-40 and the over- residuals obtained from using the second-degree polynomial
40 years old categories, this ratio was lower than one, and the ones related to the two-segment regression. The
indicating that a continuous function (a line or a horizontal line located at the value 1 represents the threshold
parabola), with no abrupt change in vertical speed, between continuous ascents (below the line) and discontin-
uous ones (above the line). In both groups, there is a
better described their uphill motion. This ascent tendency for faster athletes to perform ascents without a
profile index (the residuals ratio) shows an inverse sudden vertical velocity change (see text).
relationship with the average vertical speed in both
athlete categories (Fig. 6). When the time of inflec-
tion, expressed as a fraction of the individual race the more discontinuous ascent occurring in an
time, is plotted against the average vertical speed, earlier phase of the race, the worse the overall
a positive trend is apparent (Fig. 7). In synthesis, performance.

6
 Skyscraper running physio-mechanics
0.7
Under 40
Furthermore, the age-related decline of anaerobic
Over 40
capacity is more pronounced than that of aerobic
0.6 power, as confirmed recently by Kostka et al. (2009).
ascent inflection point (t/trace)

This evidence partially explains the lower participa-

0.5 tion of master athletes in short run-up races (less tall
skyscrapers).
0.4
The results from the present study of the time
constant (t), despite their predictable increase with
0.3
age in both genders, should be considered with
0.2
caution. Because Wilkie’s model deals with mechan-
ical power, the time constant is only loosely related
0.1 to the transient on-phase of the ‘‘global’’ oxygen
uptake, reflecting preferentially the mechanical and
0.0 the ‘‘local’’ biochemical transient from rest to max-
0.2 0.3 0.4 0.5 0.6 0.7 imal exercise levels. Also, Wilkie (1980) warns about
average vertical speed (m/s) the applicability range of his model, from 40 s to
Fig. 7. The ascent inflection time, calculated by the algo- 10 min, which almost excludes the influence of the
rithm for the two-segment regression and expressed as a small t obtained from his study and the present
fraction of the total race duration, is plotted against the investigation (ranging from 63 to 835 s). The present
average vertical speed for seven under-40 (solid squares) and
six over-40 (open squares)-year-old athletes.
t values are even smaller than Wilkie’s results, and
this could depend on the difference between the two
exercise forms: more muscle mass is activated in stair
Discussion ascending than in pedalling on a cyclo-ergometer. In
any case, obtaining t values much smaller than the
As indicated in the ‘‘Analysis of world records’’ shortest event measured suggests that those estimates
section, the physiology and biomechanics of run- have to be considered as ‘‘extrapolated’’ results, with
ups are supposed to be quite straightforward. Eighty little effect on the overall dynamics. However, a
percent of the total mechanical/metabolic work or mechanical delay in the off–on transient is expected
power is determined by the increase of body potential to increase with age, just as it occurs for the on-
energy inherent to ascending upstairs at a gradient of response time of the oxygen uptake (on average from
about 50%, the rest being caused by the cost of 25.0  3.4 to 42.0  5.1 s, from 24.2  1.8 to
running in circles between flights of stairs (15%) and 69.4  1.7 years), (Di Prampero, 2003; DeLorey et
by the acceleration of upper and lower limbs with al., 2004, 2005; Sabapathy et al., 2004; Gurd et al.,
respect to the BCOM (5%). However, different from 2008).
the high consistency of records of running on a level The reliability of the three model parameters de-
surface (Minetti, 2004) or run-up records in the same pends both on the number of athletes competing in a
building across many years (Fig. 4), the maximum given age range (general reliability) and on the atten-
estimated performance in different buildings, in dance to short duration events (B and t reliability). In
terms of mechanical work, seems quite scattered this respect, we observed, as mentioned, that old
(solid circles in Fig. 1). This may be due to building athletes tend not to compete in short run-up races.
heterogeneity (quoted above), including the variable Down to the scale of individual athletes, this study
length of the level approach to stairs, the variable shows that a digital altimeter and a HR monitor,
number of floor heights (and number) and the actual both with logging capabilities, are enough to capture
vertical distance travelled, which is not always prop- and partially explain the race result (the lactate
erly reported. concentration at the arrival was positively but not
Despite of this variability, which will never be significantly correlated to the average mechanical
alleviated by race field standardization in this sport, power). The time course of the ascent deserves
information on the aerobic and anaerobic profile of some attention, because many of the investigated
the competitors can be obtained. The decay of the runners showed a biphasic profile: the first part of the
physiological functional human capacity due to the ascent was performed at a higher vertical speed than
aging is well known (Tanaka & Seals, 2003). In the the rest of the race, with an abrupt point of inflection
present work, published race results, ranked by age (Fig. 5, column A). A more indepth analysis seems to
ranges and gender, together with the model proposed indicate that best runners (higher average vertical
by Wilkie (1980), allow to show the decay of speed) show a more uniform ascent profile, without
the (mechanical equivalent of the) aerobic and anae- an inflection point (Fig. 6). Also, an early inflection
robic capacity of male and female athletes with age point seems to be negatively related to the overall
(Fig. 3). performance. Only one athlete decided to change his

7
Minetti et al.
propulsive contribution of the upper limbs (lower
muscle mass).
Also, it is possible that less experienced athletes
tend to initially outperform to gain a leading position
in their battery, being convinced that the progression
along narrow stairs could be slowed down when
moving within a crowd.
These aspects certainly deserve further investiga-
tion. The still-limited specialization of athletes
and the competition ground heterogeneity, men-
tioned above, are reflected by the wide scatter of
current world records. A strict standardization of
ascent characteristics in future run-up events is out of
question, but in a few years’ time the records of this
relatively novel sport activity could display much less
variability due to a more focused selection of athletes
Fig. 8. Sketch of the variables involved in the calculation of and more specialized/specific training regimes.
leaning angle of the body during running in a circle (see
Appendix).
Perspectives
This investigation represents an applied physiology
profile in the second half of the ascent, by increasing study where a novel motor activity is ‘‘dissected’’ into
his velocity. Probably, he saved some energy for the its metabolic and mechanical aspects and their de-
last part of the race, but he did not win the race terminants. Also, the analysis of run-up world re-
despite having increased the average vertical speed, cords and experiments conducted during two of
underlining our hypothesis that best performance is those races allowed to reach conclusions both at
associated with a uniform ascent profile (Fig. 7). the general and at the individual level.
These observations refer to the individual strategy The novel approach to performance estimation
of conduct during the race. As in many other suggests that vertical speed and HR logging are the
competitive sports, the final result depends both on keys to explain and predict performances in this
the size of the engine and on a proper management of recent sport discipline, on which there is no scientific
energetic resources during the whole event. It is literature at present. The developed algorithm to
evident from column B in Fig. 5 how it is possible identify the inflection point in the ascent speed could
to climb the building in a shorter time by maintaining be incorporated, together with a high-resolution
a very steady heat rate and mechanical power output digital altimeter, into portable HR monitors. The
(the slope of the ascent profile), rather than starting resulting device would guide athletes during the
at a pace that cannot be maintained throughout the training sessions and even during races for the
entire race (column A). optimization of metabolic resources.
Our study is the first investigation of this new sport
activity; therefore, we cannot draw ultimate conclu- Key words: run-up race, mechanical power, digital
sions on the true determinants of the observed points altimeter, heart rate, aging, Wilkie’s Model.
of inflection in the ascent profile, which represents an
index of power switch, a phenomenon not easily
detectable in other sport disciplines. Speculative
Acknowledgements
factors include (1) negative effects of anaerobic
metabolism, particularly lactate accumulation and We wish to thank Garmin Italy for providing an additional
low tolerance, on the aerobic pathways, as partially portable GPS, including an HR monitor and a digital alti-
meter, and Thomas Dold, Flavio Ghidini, Marco Zanchi and
witnessed by the initial HR overshoot and (2) central Mario Poletti for providing their own altimetric and HR data
or peripheral fatigue, presumably also affecting the after the race.

References
Cavagna GA, Kaneko M. Mechanical genation at the onset of heavy-intensity during the transition to moderate-
work and efficiency in level walking and exercise in young and older adults. J intensity exercise. J Appl Physiol 2004:
running. J Physiol 1977: 268: 467–481. Appl Physiol 2005: 98: 1697–1704. 97: 165–172.
DeLorey DS, Kowalchuk JM, Paterson DeLorey DS, Paterson DH, Kowalchuk Di Prampero PE. Factors limiting
DH. Adaptation of pulmonary O2 JM. Effects of ageing on muscle maximal performance in humans. Eur J
uptake kinetics and muscle deoxy- O2 utilization and muscle oxygenation Appl Physiol 2003: 90: 420–429.

8
 Skyscraper running physio-mechanics
Gurd BJ, Peters SJ, Heigenhauser GJ, (anaerobic) in man. J Appl Physiol men. Resp Physiol Neurobiol 2004:
LeBlanc PJ, Doherty TJ, Paterson DH, 1966: 21(5): 1662–1664. 139: 203–213.
Kowalchunk JM. O2 uptake kinetics, Mc Fadyen BJ, Winter DA. An Saltin B, Astrand PO. Maximal oxygen
pyruvate dehydrogenase activity, and integrated biomechanical analysis of uptake in athletes. J Appl Physiol 1967:
muscle deoxygenation in young and normal stair ascent and descent. J 23(3): 353–358.
older adults during the transition to Biomech 1988: 21(9): 733–744. Tanaka H, Seals DR. Invited review:
moderate-intensity exercise. Am J Minetti AE. A model equation for the dynamic exercise performance in
Physiol Regul Integr Comp Physiol prediction of mechanical internal work Masters athletes: insight into the effects
2008: 294: R577–R584. of terrestrial locomotion. J Biomech of primary human aging on
Jones RH, Molitoris BA. A statistical 1998: 31: 463–468. physiological functional capacity. J
method for determining the breakpoint Minetti AE. Passive tools for enhancing Appl Physiol 2003: 95(5): 2152–2162.
of two lines. Anal Biochem 1984: 141: muscle-driven motion and locomotion. Teh KC, Aziz AR. Heart rate, oxygen
287–290. J Exp Biol 2004: 207: 1265–1272. uptake, and energy cost of ascending
Kostka T, Drygas W, Jegier A, Zaniewicz Minetti AE, Ardigò LP, Saibene F. and descending the stairs. Med Sci
D. Aerobic and anaerobic power in Mechanical determinants of the Sport Exerc 2002: 34(4): 695–699.
relation to age and physical activity minimum energy cost of gradient Wilkie DR. Equations describing power
in 354 men aged 20–88 years. running. J Exp Biol 1994: 195: input by humans as a function of
Int J Sports Med 2009: 30(3): 211–225. duration of exercise. In: Cerretelli P,
225–230. O’Connel ER, Thomas PC, Cady LD, Whipp BJ, eds. Exercise bioenergetics
Larsen AH, Søresen H, Puggaard L, Karwasky RJ. Energy costs of and gas exchange. Amsterdam:
Aagaard P. Biomechanical simulated stair climbing as a job- Elsevier/North Holland, 1980: 75–80.
determinants of maximal stair related task in fire fighting. J Occup Yu B, Kienbacher T, Growney ES,
climbing capacity in healthy elderly Med 1986: 28(4): 282–284. Johnson ME, An KN. Reproducibility
women. Scand J Med Sci Sports 2009: Sabapathy S, Scheider DA, Comadira G, of the kinematics and kinetics of the
19(5): 678–686. Johnston I, Morris NR. Oxygen uptake lower extremity during normal stair-
Margaria R, Aghemo P, Rovelli E. kinetics during severe exercise: a climbing. J Orthop Res 1997: 15(3):
Measurement of muscular power comparison between young and older 348–352.

Appendix
HR relationship obtained previously in the labora-
Because of the lack of relevant literature on this topic, tory. To calculate the metabolic
the metabolic energy consumption involved in run- . cost of transport for
running in circles, the net VO2 was divided by the
ning in circles, needed to approximate the total work speed of the BCOM. This speed is lower than the one
of skyscraper running, has been obtained by the (of the feet) along the figure-of-8 path according to
following procedure. the different leaning of the body, necessary to coun-
In two trained male subjects, the relationships teract the centrifugal effect of running in circles.
between heart rate (HR monitor S810i, Polar, Fin- It can be demonstrated that the relationship bet-
land, 1 beat resolution) and oxygen consumption ween leaning angle (a, deg.), circle radius (rf, m) and
(Vmax, Cardinal Health, USA) were assessed by level tangential speed along the path (vf, m/s) is
running on a treadmill at increasing speed (from 0.8
to 2.8 m/s, step 0.33 m/s). After 8 min of warm-up at r2f g
slow pace, each speed lasted 4 min, at the end of tan a ¼
ðrf  l cos aÞv2f
which measurements took place. The average values
of the two variables at all speeds were correlated using where (m) is the average height of the BCOM.
a type II linear regression (R2 5 0.957, n 5 7). We used an equation graphing software (Grapher,
Two weeks later, the same subjects, equipped with Apple Inc., Cupertino, California, USA) to calculate
HR monitors, underwent an experimental session that at the measured vf range (1.64–2.49 m/s), as
involving running in circles. On the flat roof of our measured, and for an rf of 1 m, as expected in the
department, two 1 m radius circles were drawn as to transition between stair ramps, a ranged from 781 to
produce a ‘‘figure-of-8’’ path. This was arranged to 681. Then, the speed of the BCOM (vcm, m/s)
prevent the inevitable dizziness after a few minutes of rcm
running along a single circle. After a period of vcm ¼ vf
practicing, subjects were asked to run for 5 min at rf
three speeds, individually selected from preliminary where rcm ¼ rf  l cos a (Fig. 8), was in the range
attempts as to involve an HR in the range of 135– from 1.37 to 1.79 m/s. Finally, the cost of running in
170 b.p.m. The stride frequencies during the prelimin- circles with a 1 m radius was found to be almost speed
ary tests were measured and later reproduced by a independent, and equal to 283.1  64.1 ml O2/
metronome. The tests, preceded by a 15-min warm-up (kg km). The mechanical equivalent work [J/(kg m)]
at leisure speed, were repeated three times, 10-min was obtained by multiplying the metabolic cost by the
intervals between them. muscle efficiency (0.25), and expressed as J/(kg floor)
From
. the average HRs measured, the correspond- . by assuming that two half-circles are normally ex-
ing VO2 values were estimated according to the VO2 / pected to be travelled for each floor of the building.

View publication stats