OBES SURG (2013) 23:167–172

DOI 10.1007/s11695-012-0753-6

CLINICAL RESEARCH

Age as a Predictive Factor of Testosterone Improvement

in Male Patients After Bariatric Surgery: Preliminary Results
of a Monocentric Prospective Study
Enrico Facchiano & Stefano Scaringi & Marco Veltri &
Jinous Samavat & Mario Maggi & Gianni Forti &
Michaela Luconi & Marcello Lucchese

Published online: 29 August 2012

# Springer Science+Business Media, LLC 2012

Abstract improvement in the sex hormone profile was more evident

Background Male obesity can be associated with symptom- in younger patients, with a statistically significant difference
atic alterations in sex hormones resulting in hypogonadism in cFT following surgery and in the raise of TT and cFT
and impaired fertility. Surgical-induced weight loss can between the groups of patients below and above 35 years.
improve the sex hormone profile in men. The aim of the At multivariate analysis, the age was the best predictive
present study is to evaluate the levels of sex hormones in factor of the postoperative variations of TT.
obese males before and after 6 months from bariatric sur- Conclusions These preliminary results confirm the general
gery. Possible mechanisms and clinical implications are also improvement in sex hormone profile in obese men after
discussed. bariatric surgery and introduce the age as a possible con-
Methods We evaluated levels of serum total testosterone tributing factor to this improvement.
(TT), sex hormone-binding globulin (SHBG), calculated
free testosterone (cFT), follicular-stimulating hormone Keywords Laparoscopy . Obesity . Bariatric surgery . BMI .
(FSH), luteinizing hormone (LH), and total estradiol (E2) Comorbidities . Sex hormones . Testosterone . Estradiol .
in 20 male patients at the baseline and 6 months after FSH . LH . SHBG . Hypogonadism . Infertility
bariatric surgery.
Results Median [interquartile range] age at the time of sur-
gery was 40.5 [27.2–46.7] years with a median [interquartile Introduction
range] BMI of 43.6 [40.9–48.7] kg/m2. The median baseline
levels of TT, SHBG, cFT, LH, and FSH were reduced; levels Obesity is in constant progression worldwide concerning up
of E2 were elevated. At 6 months from surgery, the median to 30 % of the population in the developed countries [1–3].
BMI dropped to 34.8 [31.7–40.5]kg/m2, TT, SHBG, cFT, This condition can be associated with different comorbid-
LH, and FSH increased, while levels of E2 decreased. The ities such as type 2 diabetes, sleep apnea syndrome, and
hypertension and with some endocrine alterations that may
E. Facchiano (*) : S. Scaringi : M. Veltri : M. Lucchese include the hormones implicated in the reproductive func-
Bariatric and Metabolic Surgery Unit, Careggi University tion in both sexes [4, 5].
Hospital,
Largo Brambilla, 3,
In male subjects, the alterations in sex hormone concen-
50134 Florence, Italy trations and actions related to obesity may result in hypo-
e-mail: enricofacchiano@[Link] gonadism and impaired fertility [6–8]. As a matter of fact,
even if the relationship between obesity-related hormonal
J. Samavat : G. Forti : M. Luconi
changes and reduced fertility is better studied in women,
Department of Clinical Physiopathology, Endocrinology Unit,
University of Florence, some reports focalize on alteration in reproductive function
Florence, Italy in males due to increased BMI [5, 7]. Obesity can affect the
normal concentrations of male sex hormones at different
M. Maggi
levels, including a reduced central production and an in-
Department of Clinical Physiopathology, Sexual Medicine and
Andrology Unit, University of Florence, creased peripheral degradation [6, 8]. However, the reduc-
Florence, Italy tion of the body fat mass through weight loss can improve
168 OBES SURG (2013) 23:167–172

the endocrine modifications induced by obesity [8–10]. An was started 1–3 months after implantation under radiologi-
effective improvement in the sex hormone profile in men cal control (upper X-ray series).
can be observed even if weight loss is induced by bariatric Laparoscopic Roux-en-Y gastric bypass (LGBP) was
surgery [4, 8, 11]. In the present prospective study, we performed by creating an alimentary and a biliopancre-
evaluated the levels of sex hormones in obese males before atic limb of 100 and 75 cm, respectively. A 15–30-ml
and after 6 months from bariatric surgery, in order to assess gastric pouch was fashioned using a 60-mm linear sta-
the effects of weight loss on endocrine changes and to pler. The gastrojejunostomy was realized in an antecolic
discuss the underlying mechanism. Possible implications antegastric fashion using a 60-mm linear stapler. Mes-
for patients' selection and their role in surgical indications enteric defects at the gastrojejunostomy site were closed
are also discussed. using running sutures to minimize the risk of internal
hernia [13].
Laparoscopic biliopancreatic diversions (LBPD) were
Methods performed using the technique described by Scopinaro et
al. [14]. A gastric distal resection leaving a stomach of 200
All male patients referred to our Unit of Bariatric and to 500 ml was realized, and then a classical Roux-en-Y limb
Metabolic Surgery on July 1, 2009 and June 30, 2010 were was built with a common limb of 50 cm and an alimentary
considered eligible for the present study. Inclusion criteria limb of 200 cm [14, 15]. Cholecystectomy was not routinely
were a BMI ≥40 kg/m2, the indication for a bariatric proce- performed.
dure to treat obesity, and no history of present or past
treatment for hypogonadism. The specific bariatric proce- Statistical Analysis
dure was chosen by a multidisciplinary team including a
bariatric surgeon, a dietitian, an endocrinologist, and a psy- A power analysis was performed using PIFACE software to
chiatrist. Only the patients operated on between January 1, define the minimal number of patients to be enrolled (n010)
2010 and September 30, 2011 giving their informed consent to obtain a statistically significant increase in post-surgery
to the surgical procedure and to the present study were total testosterone as reported in the literature [11]. Data were
considered. The baseline evaluation (T1) was performed at normally distributed according to the Kolmogorov–Smirnov
hospital admission for the bariatric procedure; the follow-up test; however, nonparametric tests as Wilcoxon and Mann–
evaluation (T2) was made after 6 months, with a tolerance Whitney U tests were used for paired and independent
of ±30 days. A sample of peripheral venous blood was sample analysis, respectively, due to the small sample size.
drawn in the morning. Pearson's correlation was used. Stepwise multiple linear
regressions were applied for multivariate analysis when
Parameters Studied indicated. A value of P<0.05 was assumed as statistically
significant. Receiver operating characteristic (ROC) curve
The following parameters were collected: analysis was performed to select the age value showing
better sensitivity and specificity for discriminating the high-
1. Anthropometric measures: height, weight, and BMI
est variation in free testosterone at 6-month follow-up. All
2. Sex hormones: serum total testosterone (TT), total estradiol
statistical analysis was performed using the SPSS version
(E2), follicular-stimulating hormone (FSH), luteinizing
18.0 software for Windows (Statistical Package for the
hormone (LH), and sex hormone-binding globulin
Social Sciences, Chicago, USA) or GraphPad Prism version
(SHBG); measured on a sample of venous blood by en-
4.00 for Windows (GraphPad Software, San Diego, CA,
zyme immunoassay (Immulite 2000, M-Medical System,
USA).
Italy). The analytical sensitivity of the essays was 0.5 nM
for TT, 55 pM for E2, 0.1 mIU/ml for FSH, 0.05 mIU/ml
for LH, and 0.02 nM for SHBG. Calculated free testoster-
Results
one (cFT) was calculated from total testosterone, SHBG,
and albumin using the Vermeulen equation [12].
Twenty obese patients were enrolled: the median age at the
Data were expressed as median [interquartile range]. baseline was 40.5 [27.2–46.7] years with a median BMI of
43.6 [40.9–48.7] kg/m2. Among these, ten patients under-
Surgical Procedures went LGBP, eight LGB, and two LBPD. At 6 months from
surgery, BMI (40.5 [27.2–46.7] to 34.8 [31.7–40.5] kg/m2,
All the surgical procedures were performed laparoscopi- P<0.0001) and E2 (from 149.5 [134.2–183.5] to 112.0 [89.9–
cally. Laparoscopic gastric banding (LGB) was performed 160.7] pmol/l, P00.002) decreased significantly, while TT
using a pars flaccida approach. Gastric banding calibration (from 8.1 [6.9–9.6] to 13.2 [10.5–18.4] nM, P<0.001), SHBG
OBES SURG (2013) 23:167–172 169

Table 1 Parameters studied at baseline and 6 months after surgery in the cohort of obese men undergoing bariatric surgery

Parameters Units Baseline levels 6 months after bariatric surgery P

Age Years 40.5 [27.2–46.7] – –

BMI kg/m2 43.6 [40.9–48.7] 34.8 [31.7–40.5] <0.0001
Total testosterone (TT) nM 8.1 [6.9–9.6] 13.2 [10.5–18.4] <0.0001
Total estradiol (E2) pmol/l 149.5 [134.2–183.5] 112.0 [89.9–160.7] 0.002
Follicular-stimulating hormone (FSH) mIU/ml 3.28 [1.95–5.05] 4.17 [3.46–6.83] <0.0001
Luteinizing hormone (LH) mIU/ml 2.70 [1.36–4.41] 3.62 [2.95–4.56] 0.048
Sex hormone-binding protein (SHBG) nM 19.0 [14.4–23.1] 39.4 [26.3–49.5] <0.0001
Free testosterone (fT)a nM 0.204 [0.180–0.274] 0.251 [0.192–0.320] NS (0.100)

Data are expressed as median [interquartile range] with P obtained by applying nonparametric Wilcoxon test for paired data
SE standard error, BMI body mass index, NS not significant
a
Free Testosterone was calculated using the Vermeulen equation from SHBG and TT values

(from 19.0 [14.4–23.1] to 39.4 [26.3–49.5] nM, P<0.0001), association between androgen variation and age retains sig-
FSH (from 3.28 [1.95–5.05] to 4.17 [3.46–6.83] mIU/ml, nificance (Table 3).
P<0.001), and LH (from 2.70 [1.36–4.41] to 3.62 [2.95–
4.56] mIU/l, P00.048) increased significantly. Also, cFT
increased (from 0.204 [0.180–0.274] to 0.251 [0.192–0.320] Discussion
nM, P00.100), but this difference failed to reach a statistical
significance when all patients were considered. However, both Men represent about 20 % of obese patients undergoing
for TT and cFT, bariatric surgery modified circulating levels bariatric surgery [16]. The anomalies of sex hormones due
from baseline values characterizing hypogonadal condition to obesity and their improvement after bariatric surgery are
(cFT <0.220 nM and TT <10.4 nM) to normogonadal levels. well studied in women, but data concerning men are still
Data of BMI, TT, SHBG, E2, FSH, LH, and cFT at baseline limited [4, 8, 11, 16–19].
and 6 months after the bariatric procedure are summarized in In physiological conditions, testosterone is mainly pro-
Table 1. Median [interquartile range] percentage of excess duced by the Leydig cells in the testes under the direct
weight loss (EWL%) at 6 months was 49.8 [29.9–65.7] %.
ROC curve analysis was used to identify the age asso-
ciated with the highest increase in cFT levels (Fig. 1). An
age below 35 identifies with a 86 % sensitivity and 83 %
specificity a cFT increase below the second tertile, with an
accuracy (area under the ROC curve) of 82 % (P<0.05).
Hence, this threshold was used for further statistical
analysis.
A significant difference in the raise of TT and cFT after
surgery (delta TT and cFT, calculated as the difference
between baseline and 6 months levels) was found between
the two groups of patients below and above 35 years,
respectively (10.54 [4.40–15.42] vs 4.15 [1.16–6.57] nM,
P00.043 for TT and 0.156 [0.063–0.177] vs 0.024 [−0.91
to 0.059] nM, P00.005 for cFT); median BMI, EWL%,
SHBG, E2, FSH, and LH did not differ significantly
between the two groups (Table 2). Moreover, delta-free
testosterone became statistically significant when consid- Fig. 1 ROC analysis displays the relationship between the true positive
ering age ≤35 (P<0.028, n07) compared to age >35 years rate (sensitivity) and the false positive rate (1-specificity) when a continue
(P<0.972, n013, paired Wilcoxon test). Regression analysis age threshold was used for discriminating the free testosterone increase
below the second tertile. The ROC curve represents the relation between
indicates a significant association between androgen level specificity and sensitivity for different thresholds of age (continuous line)
variation and age (Fig. 2a, b). When delta BMI was introduced compared to the diagonal reference line (dot line, sensitivity01-specificity).
as a further covariate in a multiple regression model, the The area under the curve represents the accuracy of the analysis
170 OBES SURG (2013) 23:167–172

Table 2 Differences in BMI and sex hormone parameters between baseline and 6 months follow-up in the two groups of patients divided according
to age cutoff of 35 years

Units ≤35 years >35 years P

Number of patients 7 13 –
Age Years 26.0 [22.0–31.0] 45.0 [40.5–50.0] <0.0001
ΔBMI kg/m2 11.7 [9.9–18.7] 8.0 [4.2–12.8] NS (0.322)
EWL % 49.8 [34.4–62.9] 49.8 [23.0–67.8] NS (0.905)
ΔTotal testosterone (TT) nM 10.54 [4.40–15.42] 4.15 [1.16–6.57] 0.043
ΔTotal estradiol (E2) pmol/l −3.0 [−77.0 to 29.0] −39.0 [−57.0 to –19.5] NS (0.552)
ΔFollicular-stimulating hormone (FSH) mIU/ml 1.82 [0.48–3.08] 0.77 [0.18–1.98] NS (0.362)
ΔLuteinizing hormone (LH) mIU/ml 0.27 [−0.38–1.99] 1.16 [−0.77–2.65] NS (0.721)
ΔSex hormone-binding protein (SHBG) nM 11.2 [1.1–43.3] 16.0 [6.9–24.7] NS (0.968)
ΔFree testosterone (fT)a nM 0.156 [0.063–0.177] 0.024 [−0.91–0.059] 0.005

Data are expressed as median [interquartile range] with P obtained applying nonparametric Mann–Whitney U test for independent data
BMI body mass index, NS not significant, EWL% percent of excess weight loss
a
Free testosterone was calculated using the Vermuelen equation from SHBG and TT values

stimulation of LH and one of the way of its peripheral concentrations of testosterone is well-known and the de-
degradation is represented by conversion to estradiol crease in androgen levels is proportional to the degree of
mediated by the enzyme aromatase, normally present obesity [7, 8, 18, 22, 23]. As a matter of fact, these
in the white adipose tissue [6–8]. Obese men, as they alterations in sex hormones can partly explain the role of
present high levels of adipose tissue, exhibit an elevated obesity in the male infertility, besides the almost rare ge-
aromatase activity with increased conversion of andro- netic factors and the potential physical detrimental mecha-
gens into estrogens, finally resulting in elevated estradi- nisms such as the increased scrotal temperature observed
ol and reduced testosterone-circulating concentrations [6, following fat increase [6].
20]. Moreover, obese males often show low levels of Alterations in sex hormones associated with obesity condi-
FSH, which directly stimulates the Sertoli cells in the tion can beneficiate from weight loss with a general improve-
testis tubules. The inappropriately low levels of LH, ment in the hormonal profile which can be observed even when
despite the reduced levels of testosterone, may be due to the the weight loss is obtained through bariatric surgery [4, 6–8, 11,
relative suppression caused by raised concentrations of estro- 16–18]. The few series published so far analyzing the weight
gens [6, 7, 21]. The hormonal modifications characterized by loss induced by different bariatric procedures found that testos-
decreased testosterone and gonadotropin levels and increased terone (both TT and cFT), SHBG, and LH generally improve
estrogen concentrations represent a hormonal profile typically after weight loss [4, 7, 11, 16–18]. Only the results about FSH
observed in obese male (hypogonadotropic hyperestrogenic remain controversial, with some authors reporting an increase
hypogonadism) [7]. The association between obesity and low [4, 11, 17] and others reporting a decrease [16, 18] after

Fig. 2 Relationship between

androgen variations (delta) after
bariatric surgery and baseline
age. Linear regression curves
are shown for TT (a) and cFT
(b), respectively. R2 00.266,
R0−0.516, P00.020 and
R2 00.375, R0−0.612, P00.004
for delta TT and delta cFT,
respectively
OBES SURG (2013) 23:167–172 171

Table 3 Multivariate analysis of testosterone and free testosterone stimulation (due to an LH increase) or a decreased periph-
changes (Δ) between baseline and 6 months follow-up in operated
eral degradation (with a consequent decrease in E2 levels),
patients
but only to an improved testicular function. This should be
ΔTT R P more evident in younger patients, maybe as a consequence
Age −0.516 0.020 of a better testicular response. The clinical consequence of
Age −0.360 NS (0.066) this observation can be that the patients who better beneficiate
ΔBMI 0.639 0.009 of the surgically obtained weight loss in term of testosterone
increase are the younger ones and this could have an important
ΔfT R P implication in patients' selection for bariatric surgery. Maybe,
Age −0.612 0.004 in the case of hypogonadism in young obese patients, partic-
Age −0.563 0.016 ularly in the presence of altered sexual function, the young age
ΔBMI 0.266 NS (0.291) (≤35 years, according to our results) should constitute an
elective element to propose bariatric surgery.
R and P coefficients are indicated. Partial and adjusted correlations are Several limitations of the present study should be recog-
indicated for the independent variables considered
nized. First, the small sample size may result in a type II
error, which could be responsible for the lack of statistical
bariatric surgery, even if these variations often fail to reach a significance for some parameters. However, the power anal-
statistical significance. ysis we performed at the beginning indicates that the num-
All the obese patients reported in our series presented ber of patients enrolled is adequate for assessing differences
reduced baseline levels of total testosterone, SHBG, free in testosterone levels, as reported in the literature [11].
testosterone, LH, and FSH and elevated levels of E2. As Another limitation is the different surgical procedures
expected, after 6 months and after a significant weight loss employed to obtain weight loss. Finally, the short follow-
had occurred, we found a significant raise in the levels of up of 6 months indicates the need for further studies.
testosterone, SHBG, LH, and FSH and an important drop in In conclusion, these preliminary results confirm the gen-
E2. These results are consistent with those reported in other eral improvement in sex hormonal profile in obese men after
series investigating the changes in sex hormones after bari- bariatric surgery, already observed in other series. More-
atric surgery [4, 7, 11, 16–18]. The amelioration of the over, we introduce the age as a possible contributing factor
sexual hormonal profile in obese men after bariatric surgery to this improvement. Our preliminary data, if confirmed by a
could have some clinical implications: the improvement of longer follow-up and in larger series, could influence the
testosterone levels after bariatric surgery, in fact, could processes of patients' selection for bariatric surgery,
result in considering secondary hypogonadism and male introducing male hypogonadism as comorbidity that
infertility associated to obesity as a real comorbidity that could ameliorate after bariatric surgery, above all in
could beneficiate from the surgical-induced weight loss. younger patients.
Moreover, as a well-demonstrated association between low
levels of testosterone and insulin resistance is clear, it can be
speculated that, in addition to the direct effect of bariatric
surgery, a general amelioration in sex hormonal profile in
Conflict of interest The authors declare that they have no proprietary,
obese male patients could reduce the risk to develop type 2 financial, professional, or other personal interest of any nature or kind in
diabetes [23–27]. This effect, however, needs to be con- any product or service that may influence the position presented.
firmed by further studies designed with this specific aim.
The most important finding of our report lies in the fact
that the patients' age at surgery strongly correlates with the
increase in testosterone levels observed at 6 months from References
surgery. When we divided our patients in two groups of age
considering a cutoff of 35 years, derived from the specific 1. Flegal KM. Epidemiologic aspects of overweight and obesity in
ROC curve analysis, we found a statistical difference in the the United States. Physiol Behav. 2005;86(5):599–602.
2. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight
testosterone raise between the two groups without observing
and obesity in the United States, 1999–2004. JAMA. 2006;295
any difference in the other parameters. In particular, the (13):1549–55.
recovery in free testosterone became statistically significant 3. Sjostrom L, Narbro K, Sjostrom CD, et al. Effects of bariatric
when considering age ≤35 years. The fact that the variations surgery on mortality in Swedish obese subjects. N Engl J Med.
2007;357(8):741–52.
in both LH and E2 did not differ between the two groups of
4. Alagna S, Cossu ML, Gallo P, et al. Biliopancreatic diversion:
age could lead us to hypothesize that the difference in the long-term effects on gonadal function in severely obese men. Surg
testosterone increase is not due to an increased central Obes Relat Dis. 2006;2(2):82–6.
172 OBES SURG (2013) 23:167–172

5. Facchiano E, Iannelli A, Santulli P, et al. Pregnancy after laparoscopic 17. Bastounis EA, Karayiannakis AJ, Syrigos K, et al. Sex hormone
bariatric surgery: comparative study of adjustable gastric banding and changes in morbidly obese patients after vertical banded gastroplasty.
Roux-en-Y gastric bypass. Surg Obes Relat, Dis. 2011;8(4):429–33. Eur Surg Res. 1998;30(1):43–7.
6. Cabler S, Agarwal A, Flint M, et al. Obesity: modern man's 18. Globerman H, Shen-Orr Z, Karnieli E, et al. Inhibin B in men with
fertility nemesis. Asian J Androl. 2010;12(4):480–9. severe obesity and after weight reduction following gastroplasty.
7. Hammoud AO, Gibson M, Peterson CM, et al. Impact of male Endocr Res. 2005;31(1):17–26.
obesity on infertility: a critical review of the current literature. 19. Pasquali R, Pelusi C, Genghini S, et al. Obesity and reproductive
Fertil Steril. 2008;90(4):897–904. disorders in women. Hum Reprod Update. 2003;9(4):359–72.
8. Rao SR, Kini S, Tamler R. Sex hormones and bariatric surgery in 20. Roth MY, Amory JK, Page ST. Treatment of male infertility
men. Gend Med. 2011;8(5):300–11. secondary to morbid obesity. Nat Clin Pract Endocrinol Metab.
9. Kaukua J, Pekkarinen T, Sane T, et al. Sex hormones and sexual 2008;4(7):415–9.
function in obese men losing weight. Obes Res. 2003;11(6):689–94. 21. Akingbemi BT. Estrogen regulation of testicular function. Reprod
10. Stanik S, Dornfeld LP, Maxwell MH, et al. The effect of weight Biol Endocrinol. 2005;3:51.
loss on reproductive hormones in obese men. J Clin Endocrinol 22. Giagulli VA, Kaufman JM, Vermeulen A. Pathogenesis of the
Metab. 1981;53(4):828–32. decreased androgen levels in obese men. J Clin Endocrinol Metab.
11. Hammoud A, Gibson M, Hunt SC, et al. Effect of Roux-en-Y 1994;79(4):997–1000.
gastric bypass surgery on the sex steroids and quality of life in 23. Tsai EC, Matsumoto AM, Fujimoto WY, et al. Association of
obese men. J Clin Endocrinol Metab. 2009;94(4):1329–32. bioavailable, free, and total testosterone with insulin resistance:
12. Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of influence of sex hormone-binding globulin and body fat. Diabetes
simple methods for the estimation of free testosterone in serum. J Care. 2004;27(4):861–8.
Clin Endocrinol Metab. 1999;84(10):3666–72. 24. Gautier JF, Choukem SP, Girard J. Physiology of incretins (GIP
13. Iannelli A, Facchiano E, Gugenheim J. Internal hernia after and GLP-1) and abnormalities in type 2 diabetes. Diabetes Metab.
laparoscopic Roux-en-Y gastric bypass for morbid obesity. 2008;34 Suppl 2:S65–72.
Obes Surg. 2006;16(10):1265–71. 25. Haffner SM, Karhapaa P, Mykkanen L, et al. Insulin resistance,
14. Scopinaro N, Gianetta E, Civalleri D, et al. Bilio-pancreatic bypass body fat distribution, and sex hormones in men. Diabetes. 1994;43
for obesity: II. Initial experience in man. Br J Surg. 1979;66 (2):212–9.
(9):618–20. 26. Stellato RK, Feldman HA, Hamdy O, et al. Testosterone, sex
15. Scopinaro N, Adami GF, Marinari GM, et al. Biliopancreatic hormone-binding globulin, and the development of type 2 diabetes
diversion. World J Surg. 1998;22(9):936–46. in middle-aged men: prospective results from the Massachusetts
16. Reis LO, Favaro WJ, Barreiro GC, et al. Erectile dysfunction and male aging study. Diabetes Care. 2000;23(4):490–4.
hormonal imbalance in morbidly obese male is reversed after 27. Patriti A, Facchiano E, Sanna A, et al. The enteroinsular axis and
gastric bypass surgery: a prospective randomized controlled trial. the recovery from type 2 diabetes after bariatric surgery. Obes
Int J Androl. 2010;33(5):736–44. Surg. 2004;14(6):840–8.