The Effects of Diet on Testosterone (Part 2): Carbohydrates and Fats

Part One of this article explained the impact of calories and dietary protein (PRO) on endogenous testosterone (T) levels. As promised, this continuation will focus on the role of dietary carbohydrates (CHO) and dietary fat on modulating T production. The role of CHO on T production is indirectly addressed when discussing the role of PRO or fat, so this will be reviewed briefly. The effects of fat on T are far more complicated and often time more confusing than the previously discussed macronutrients. To facilitate an understanding of the links between dietary fats or lipids and T, several tables will be presented. An explanation will accompany each table and key references will be reviewed. The article ends with an application of the information to the design of a dietary strategy to either maximize or minimize T levels.

Dietary Carbohydrate Intake & Testosterone

Dietary carbohydrates can influence the metabolism of a variety of chemicals. When fat is held at approximately 20% of caloric intake, CHO may elevate T levels (1). Part One of this article discussed that while this may be true, there is also a corresponding increase in sex hormone binding-globulin (SHBG). Anderson et al (1) compared the effects of a higher PRO diet versus a higher CHO diet on T levels. Part one discussed the data on the high protein diet. The higher CHO diet contained approximately 2450 kcals/d, 70% CHO, 10% PRO, 20% fat. This provides 429 g/d CHO, 62 g/d PRO, and 55 g/d fat. The seven men in this study had a range of body weights from 64-72 kg. If a mean of 68 kg is assumed, then these subjects were taking in .91g PRO/kg BW or slightly higher than the RDA of .8g/kg BW. This point is made because most people take in more protein than this on a daily basis.

Now let’s get back to the T and SHBG issue. The interaction between T and SHBG is important to consider. About 44% of total T is bound to SHBG and is called SHBG-T. If T increases more than the SHBG-T fraction does, then the biological actions of T will be greater because more of it will be available to bind to muscle and other tissues’ receptors. If T increases less than SHBG-T fraction, then the biological actions of T will decrease because less of it will be available to bind to muscle and other tissues’ receptors. Anderson et al did not measure SHBG-T. The study did measure total T and SHBG. It can be seen from their data, that T increases less than SHBG did on the higher CHO diet with a ratio of 7:1 (CHO:PRO). The T values were 16.2 ± 1.2 nmol/L. This was a 28% increase over the high PRO diet and the range of increases in the subjects was from 10-93%. Assuming that the SHBG-T fraction remained at 44% of T, then the amount of T that was bioavailable would be about 9.1 ± .66 nmol/L. Compared to the amount of bioavailable T on the high PRO diet, there is an additional 1.9 ± .21 nmol/L of bioavailable T.

Also keep in mind that this same type of diet increases the ability of the liver to reduce T to 5a – reduced hormones (ie androsterone) (2), which may or may not be something you want (depending on the study you read). However, this is especially important for steroid and prohormone users because a higher CHO diet may increase the conversion of the exogenous T to androsterone. This is not to say that diets with higher CHO than PRO will cause this to occur. What this means is that very high CHO:PRO ratios like 7:1 or greater may not be the healthiest way to go, based upon direct and indirect evidence that androsterone is linked to acne and prostate disorders.

The effects of CHO on T were just discussed while fat was kept constant in the diet at about 20% of calories. When PRO is kept constant in the diet, higher CHO may actually lower T (8). Hamalainen et al (8) compared the effects of a dietary intervention on the hormone levels of 30 men. PRO intake was fairly consistent while the CHO was increased from 45% to 56% of calories for six weeks, and then decreased to 47% for six weeks. Fat intake was correspondingly decreased from 40% to 25%, and then increased to 37%. During the higher CHO period, T and fT decreased significantly. However, this study was difficult to interpret because dietary fibers, like pectin from fruit or bran from wheat, and fatty acids, like saturated fatty acids or polyunsaturated fatty acids, can also have an impact on T production. In the Hamalainen et al study, they also changed the fatty acid ratios of the diets. Perhaps the ratio of fatty acids, as opposed to the amount of CHO or fat, had a larger impact on T production. Extrapolating this further, maybe it is not the amount of CHO or the CHO:PRO that influences T production, but the ratio of CHO to a particular fatty acid, or some other nutrient interaction (ie PRO to fatty acid or ratio of fatty acids).

Correlation Studies Between Dietary Fat Intake and Testosterone Levels in Men

Fat has received tremendous attention over the last few years and has been linked to improved performance and favorable body composition alterations in the lay journals, despite a lack of convincing scientific data. The relationship between dietary lipids and T is important in order to understand the role that fat may have in improving performance, altering body fat, or preventing/initiating disease.

One of the reasons why the scientific data has not been clear in explaining the role of dietary fat on T levels is a difference in study designs.Table 1 displays the data and results from several studies that compared T, free testosterone (fT), and/or SHBG levels with total fat or types of fatty acids in the diet. Data is listed as the mean values (when available). Correlation studies, while very common, are far from complete. They don’t explain if dietary fat or some fraction, like polyunsaturated fatty acids (PUFA), affects T, rather they only state if there is a relationship between one event and another. The relationship can be positive and an example of this is reference 19 from Table 1. From the results column the code FCT is listed in the results column. FCT means that as the percentage of calories from fat, grams of saturated fat, and grams of monounsaturated fat (MUFA) increased in the diet, there was also a corresponding association with higher T levels. This study was done with resistance trained males and is the most applicable from all of the above studies. The scope of this article precludes an in-depth analysis of each study and the associated design flaws. Most important is to cite the common findings. From Table 1, several relationships can be seen. Subjects consuming vegetarian diets have demonstrated higher SHBG levels (3, 13), lower T levels (12), and lower levels of available T (3). One flaw with many of these studies is isolating the impact of fat on the diet as opposed to fiber, which is also much higher in vegetarian-type diets. Another problem with correlational studies is that they don’t tell you what happens when subjects are switched from one type of diet to another. Unfortunately studies sometimes contradict each other. For example, Bishop et al (4) examined the role of dietary nutrients on sex hormone differences between monozygotic twins (identical twins). The investigators found an inverse (or negative) relationship between dietary fats and T. Volek et al (17) however, found a positive relationship between dietary fat and T. This further demonstrates the problem of reading the scientific literature and making sense of all the information.

Table 1: Correlational Studies Between Dietary Fat Intake and Testosterone Levels in Men
Ref	#	Age	Group	Fat	SFA	PUFA	P:S	CHO	PRO	Fiber	Calories	Results
#	Subj.	years		%	%	%	ratio	%	%	g/d	kcals/d
3	14	27	O	36.8	14.8	4.1	0.28	48.3	16.6	14	2620
	15	29	V	16.7	9	6.3	0.7	58.3	12.6	38	2700	HS, LAT
4	144	34.5	MZ	39.2	NG	NG	NG	41.2	18.1	6.5	3212	FINVT
5	12	33	M	NG	NG	NG	NG	NG	NG	NG	NG	NC
	34	26	NE	NG	NG	NG	NG	NG	NG	NG	NG	NC
	23	69	VM	NG	NG	NG	NG	NG	NG	NG	NG	NC
	15	73	NVM	NG	NG	NG	NG	NG	NG	NG	NG	NC
7	1241	54.4	MAM	NG	NG	NG	NG	NG	NG	NG	NG	NC
12	12	56.5	SV	34	10.5	>5.7	>.54	54	14	37	2667	LT
	10	55.3	SNV	38	12.6	>5.9	>.47	45	18	23	2292
	8	55.8	NV	40	11	>5.8	>.53	39	12.7	20	2617	HT
13	18	41.4	VEG	32.6	6.2	10.3	1.66	54.4	11.5	56.6	2581	HS
	22	40.3	O	38	12.9	7.2	0.60	43	14.4	35.1	2605
19	12	23.8	RTM	23.0	6.8	4.1	0.65	56.0	20.0	21.3	2366	FCT, PSLT
CHO = Carbohydrates; FCT = % fat, SFA, & MUFA correlated with higher T levels; TINVT =
Fat inversely correlated to T; HS = Higher SHBG levels correlated with V & VEG diets; HT =
Higher T; LAT = Lower available T; LT = Lower T; M = Macrobiotic; MAM = Middle-aged men;
MZ = Monozygotic twins (72 pairs); NC = No correlation (macronutrients vs T or fT); NE =
Normal eaters; NV = Non-Seventh Day Adventist and Non-vegetarians; NVM = Nonvege-
tarian monks; O = Omnivores; PRO = Protein; P:S = PUFA:SFA ratio; PSLT =PUFA:SFA ratio
correlated with lower T levels; PUFA = Polyunsaturated fatty acids; RTM = Resistance trained
males; SFA = Saturated fatty acids; SNV = Seventh Day Adventist Nonvegetarians; SV =
Seventh Day Adventist Vegetarians; T = Testosterone; V = Vegetarians; VEG = Vegans; VM
= Vegetarian monks

Acute Effects of Dietary Fat on Testosterone

A better study design than a correlational study to determine the effects of manipulating dietary macronutrients is a randomized cross over, double-blind study. Cross over means that every subject experiences all of the different dietary treatments. By randomizing the order, the effect of one diet on another is avoided (this is called order effect). Double-blind means that the subjects, the people working with the subjects, and the people tracking the data are all unaware of the treatment conditions. This is very difficult to do with feeding studies, so in most cases a double-blind approach is not used. Therefore, in most studies, the subjects and/or the researchers know what the treatment conditions are. One way the researchers avoid this problem is to offer milk shakes that taste the same, but, in fact, have different macronutrient compositions. While this may be acceptable to study the acute effects of more or less fat in a meal, this would not work for chronic studies. After all, could you drink the same milkshake all day long for weeks and weeks, or worse yet eat some type of engineered food product not knowing what was inside?

Acute studies examine the effects of different treatments within the hours or days after the dietary manipulation. In general, the subjects are given different types of diets and the results of each diet are compared. This is one way to look at the effects of a particular nutrient on hormone levels or blood glucose levels, for example. Table 2 presents the tabulated data from two short term or acute studies.

Table 2: Acute Effects of Fat on Testosterone in Men

Ref

#

Age

Diet

Fat

SFA

PUFA

P:S

CHO

Pro

Fiber

Calories

Time

Results

years

%

%

%

ratio

%

%

g

kcals

14

8

23-35

C

0

0

0

0

0

0

0

<1

4 hours

NC T & fT

HF

57.0

NG

NG

NG

34.0

9.0

NG

795.2

4 hours

DT & fT

LF

1.2

NG

NG

NG

73.3

25.5

NG

797.0

4 hours

NC T & fT

18

4

25.5

N

40.0

NG

NG

NG

45.0

15.0

NG

~2750/d

3 days

NC T & fT

CHO

30.0

NG

NG

NG

55.0

15.0

NG

~2750/d

3 days

NC T & fT

PF

60.0

NG

NG

NG

15.0

25.0

NG

~2750/d

3 days

NC T & fT

C = Control diet; CHO = Carbohydrate rich; DT&FT = Decreased T & fT; HF = High fat diet;

LF = Low fat diet; NC T & fT = No significant changes in T or fT; NG = Not given; PF = Protein

& fat rich

In one study (14), the effects of high fat (HF) and low fat (LF) meals on T levels were compared. The subjects were given a lemon-lime artificially sweetened beverage and the hormonal responses served as a control (C) for the other meals. A HF liquid meal containing about 795 calories and made up of 57% fat (50.4 g fat), 9% protein (17.9 g PRO), and 34% carbohydrate (67.5 g CHO) was given on another occasion. The third or final liquid meal (LF) consisted of 797 calories made up of 1.2% fat (1 g fat), 25.5 % PRO (51 g PRO), and 73.3% CHO (146 g of CHO). The C and LF meals did not effect luteinizing hormone (LH), T, fT or dihydrotestosterone (DHT) levels. The HF meal decreased T and fT up to 4 hours post ingestion compared to the other liquid meals without affecting any of the other hormones.

There are some problems with this study, however. It was not double-blind, the treatments were not randomized, it used a small sample size of eight, and while the subjects were instructed to fast, no data was offered to confirm this, like blood sugar levels. The study also did not look at the possible mechanisms by which the HF diet lowered T and fT levels.

It has been proposed in the literature that fatty acids may bind SHBG. If this is true, then after the fat is broken down from a high fat meal, a corresponding increase in blood fatty acid levels would occur, and less SHBG is available to bind with T. This would then increase the percentage of fT in the blood. However, since the percentage of fT in this study did not change (the total amount decreased, not the percentage of total T), this could not have occurred. The researchers do offer that the only way that the HF meal could have affected T/fT levels was either by increasing the clearance rate or decreasing the production rate. The clearance rate would be determined by the rate of uptake by tissues, the rate of T and fT metabolized by the liver, and the rate of excretion by the kidneys. While fatty acids do attach to T and fT inside the body, there is no data to say that this increases uptake into tissues like skeletal muscle or that the event could occur within four hours post-meal ingestion. It would be unlikely that the fatty acids from the meal could affect the liver enzymes involved in T or its fractions so soon. It is possible that ketones produced from the breakdown of the fatty acids could cause the renal tubules to excrete more T and fT. But this is unlikely due to the fact that the subjects were not in a glycogen-depleted state and there were PROs and CHOs in the meal. This leaves decreased production of T and fT as the most likely reason for the drop in these hormones. Again, this is only speculation at this point since the study did not examine the possible causes for the decrease in the hormones.

Chronic Effects of Dietary Fat on Testosterone

The chronic studies presented in Table 3 report the effects of 2 or more weeks of dietary manipulations on testosterone levels. A decrease in dietary fat has been shown to decrease total T (8, 11, 15) and fT levels (8, 16) or not affect T levels (17). Approaching this from the other direction, an increase in dietary fat has been shown to decrease total T (11), and increase (16) or decrease fT levels (6). It’s not necessary to review all the studies to try to explain the differences in results. However, notice that from the Table 3, most studies compared vegetarian-type diets to western-type diets. This presents several problems when trying to explain the hormonal responses from the dietary manipulations. The first is that other dietary factors were altered in addition to fat intake. These included fiber content and the presence of various phytonutrients like flavonoids, isothiocyanates, etc. The main point is that there are many factors that can determine the effects of dietary fat on T levels. Most studies did not even report the amounts of fatty acids in the subjects’ diets, let alone the content of phytonutrients, so these factors were most likely not controlled for. Furthermore, differences in the length of the treatments (2 weeks vs. 10 weeks), lifestyles of the subjects (active vs. sedentary), and calorie loads (2800 vs. 4374) are additional examples of factors that can impact the results.

Table 3: Chronic Effects of Fat on Testosterone in Men

Ref

#

Age

Group

Diet

Fat

SFA

PUFA

P:S

CHO

Pro

Fiber

Calories

Time

Results

years

%

%

%

ratio

%

%

g/d

kcals/d

6

43

19-56

HM

LFHF

18.8

4.4

5.7

1.30

67.5

17.1

61.2

3179

10 weeks

HFLF

41.0

14.7

8.8

0.60

45.3

14.8

26.4

3155

10 weeks

IUT, DfT

8

30

40-49

HNO

B

40.0

23.0

3.0

0.15

45.0

14.0

NG

3322

2 weeks

I

25.0

8.0

10.0

1.22

56.0

18.0

NG

2811

6 weeks

DT, DfT

S

37.0

21.0

3.0

0.15

47.0

14.0

NG

3346

6 weeks

9

4

>45

WNAM

CD

40.0

NG

NG

NG

NG

NG

NG

2800

4

VD

30.0

NG

NG

NG

NG

NG

NG

2800

3 weeks

DUT

10

16

49

WNAM

CD

40.0

NG

NG

NG

NG

NG

NG

2800

VD

30.0

NG

NG

NG

NG

NG

NG

2800

3 weeks

DUT

18

45

BNAM

CD

40.0

NG

NG

NG

NG

NG

NG

2800

VD

30.0

NG

NG

NG

NG

NG

NG

2800

3 weeks

NSD

21

43

BSAM

CD

30.0

NG

NG

NG

NG

NG

NG

2800

WD

40.0

NG

NG

NG

NG

NG

NG

2800

3 weeks

IUT

11

11

49

WNAM

CD

40.0

NG

NG

NG

NG

NG

NG

2800

IUT

VD

25.0

NG

NG

NG

NG

NG

NG

2800

3 weeks

DT

13

45

BNAM

CD

40.0

NG

NG

NG

NG

NG

NG

2800

VD

25.0

NG

NG

NG

NG

NG

NG

2800

3 weeks

DT

20

43

BSAM

CD

25.0

NG

NG

NG

NG

NG

NG

2800

DUT

WD

40.0

NG

NG

NG

NG

NG

NG

2800

3 weeks

IUT, DT

15

8

22.5

ET

H

32.0

NG

NG

0.43

51.5

14.0

47.0

4111

4 weeks

M

28.7

NG

NG

0.49

57.2

13.9

47.0

4183

6 weeks

VD

27.4

NG

NG

1.14

57.9

14.7

98.0

4374

6 weeks

DT

16

6

34

NM

HF

>100g

NG

NG

NG

NG

NG

NG

NG

2 weeks

IfT, DS

LF

<20g

NG

NG

NG

NG

NG

NG

NG

2 weeks

DfT, IS

CP

NG

NG

NG

NG

NG

NG

NG

NG

4 weeks

NSD

17

21

57

CVD

B

NG

NG

NG

NG

NG

NG

NG

NG

I

<10.0

NG

NG

1.24

77.0

13.0

>35.0

NG

26 days

NSD

B = Baseline; BNAM = Black North American men; BSAM = Black South African men; CD = Control diet;

CP = Control period; CVD = subjects in this study had coronary heart disease, hypertension, &/or diabetes;

DfT = Decreased fT; DS = Decreased SHBG; DT = Decreased T; DUT = decreased urinary T; ET =

Endurance trained; H = Habitual diet; HF = High fat diet; HFLF = High fat low fiber; HM = Healthy men;

HNO =Healthy non-obese men; I = Intervention; IfT = Increased fT; IS = Increased SHBG; IT =Increased T;

IUT = Increased urinary T; LF = Low fat diet; LFHF = Low fat high fiber; M = Mixed diet; NG = Not given;

NM = Normal men; NSD = No significant differences; S =Switchback; VD = Vegetarian diet; WD = Western

diet; WNAM = White North American men

All the Evidence Not In Yet

It has been speculated that the ratio of fatty acids may have some role on whether or not dietary fat increases or decreases T levels. A positive relationship between saturated fatty acids and monounsaturated fatty acids with T levels has been reported previously (19). The same data also describes a negative (or inverse) relationship between polyunsaturated fatty acids and T levels. These relationships between dietary fat components and T have also been supported by a study on eight men randomly assigned and crossed over from a vegetarian diet to a mixed-meat diet that was isoenergetic (15). About 28% of the calories were from fat. The vegetarian diet had a polyunsaturated fatty acid to saturated fatty acid ratio (P:S) > 1, while the mixed-meat diet had P:S of about .5.

In a 1996 study, forty-three men were exposed to a high-fat, low-fiber diet for 10 weeks and a low-fat, high-fiber diet for 10 weeks in a cross over design (6). Total T and fT did not change significantly. SHBG-bound T was higher on the high-fat diet, which does not agree with another study (16). The researchers claimed this might have been due to within-person variations of plasma testosterone levels.

Another important finding was that urinary excretion of T was much greater on the high-fat, low-fiber diet (6). Other studies have shown that on higher fat diets, urinary excretion of T is increased (10, 11) while vegetarian type diets may decrease the urinary excretion of T (9, 10, 11). This is an important point to consider in evaluating the level of T bioactivity in the body. If blood levels of T elevate and the excretion rate of T also elevates there may not be a net bioactive effect of T. However, if blood levels of T remain the same and T excretion decreases, that may signal a net bioactive effect of T in the body. While it is difficult to say if a higher fat or lower fat diet would be better for increasing the bioactivity of T, it does appear that higher fat and lower fiber-type diets are associated with greater excretion of T. An increase in the urinary excretion of T combined with an elevation of T levels in the blood may indicate that the net T production is greater. The implication is that cells may have an increased opportunity to be exposed to T. Alternatively, perhaps it is the result of some type of self-regulating mechanism that the body maintains to keep endogenous levels in check.

There are many more studies in the literature. The intent was to expose the reader to all the different possible interactions and the complexity in trying to control for all areas just to determine the role of fat on androgen production. Other studies have examined the effects of different fatty acids on testicular cell membranes and T levels after supplementation fatty acid supplementation. The results do not support one another and only point to the fact that dietary fat plays a role in modifying T production, but that role is still unclear.

Designing A Diet to Maximize Testosterone Levels

Remember, it is the bioactive fraction of total T that is important. This fraction consists of fT and albumin-bound T. Fasting suppresses T production and small amounts of either PRO or CHO do not reverse the suppression. Diets with a PRO intake greater than the CHO intake lower total T levels, and may actually decrease the bioactivity of T in the body. Higher CHO diets (70% or more from CHOs) may increase T levels, but they also affect the metabolism of T as well. While the role of fat is not entirely clear, saturated fat and cholesterol are closely linked to higher levels of T and PUFAs have some modifying role.

So, what is the best type of diet to follow if your only concern is to increase T levels and make more of it available to the body for the purpose of improving lean body mass and/or performance? It would seem that CHO intake must exceed PRO intake by at least 40% to keep the bioactive fraction of T high. Fat intake should be at least 30%, saturated fat needs to be higher than PUFA, and fiber intake needs to be low. A sample diet would have roughly the following calorie breakdown: 55% CHO, 15% PRO and 30% fat. On the other hand, what if you wanted to lower your T levels in order to minimize cardiovascular disease risk factors and/or hormone-dependent cancer risks? Then a diet with more protein, more fiber, a fat intake below 25%, and a P:S ratio of 1 or higher would be a more prudent choice. The breakdown of this sample diet would be about 50% CHO, 30% PRO and 20% fat. The problem with using percentages, however, is that people with high calorie needs will most likely take in far more protein then they need. Another strategy is to keep protein intake the same (ie 1 gram per pound of BW) and then play around with the fiber, SFA:PUFA ratio, CHO, and total fat contents of the diet. Antioxidants are important additions when trying the higher fat diets. Keep in mind there are many factors that affect T production and they interact in a complex and seemingly unpredictable fashion. We invite feedback and will respond to all questions, comments, etc. Several readers have mentioned the idea of cycling a diet that maximizes T and then switching back to a healthier type of diet.