Re: Weight loss, or Weight loss Diet? Your local actuary wants to know......

[Guest guest]

The question is, does voluntary weight loss cause ill effects? Well, it will turn on your doctor for more tests, maybe better health care, because she won't believe it's voluntary. I think we have to lose weight because the heart muscle ages, doesn't perform as well. As long as I can add weight easily, it implies to me, my heart is healthier, suffers less from athero perhaps.

I just can't see where I'd get to 100 yo weighing 178#, so I wonder about losing at a controlled rate, say 2# per year for 30 yrs. I'd be 118#, not bad for a 100 yo, IMO.

Forcing to a lower weight earlier, or not - I see no data for that. Especially since it's hard to do, and anorexia looms, if I go to far.

Regards.

----- Original Message -----

From: T

Sent: Sunday, July 24, 2005 3:21 AM

Subject: [ ] Weight loss, or Weight loss "Diet"? Your local actuary wants to know......

Hello out there in Beaver Dam!

"After controlling for age, medical, and lifestyle factors, both men and women had higher mortality rates over a 10+ year period for increasing categories of weight loss......"

"Persons on weight loss diets within the year prior to baseline did not have increased mortality with increasing weight loss. CONCLUSION: The strong association between weight loss (likely involuntary) and mortality may be a useful way of estimating overall risks to longevity in populations."

I think this type of research is important, if only for highlighting the limitations of epidemiologic research. You have to carefully control for this type of thing, or, invariably, you will get the "frail" thrown in with the the folks innocently following the Schwarzbein Principle.

=-=-=-=-=-=-=-=-=-=-=--=-=-=

Ann Epidemiol. 2005 Aug;15(7):483-91.

Related Articles,

Links

Associations with weight loss and subsequent mortality risk.Knudtson MD, Klein BE, Klein R, Shankar A.From the Department of Ophthalmology and Visual Sciences, University of Wisconsin Medical School, Madison, WI, USA.PURPOSE: Studies have shown a high prevalence of weight loss in older adults is associated with an increased risk of death. We investigated this in a population-based study. METHODS: Persons living in Beaver Dam, Wisconsin, participated in a baseline examination between 1988 and 1990 (n=4926). A medical examination and standardized questionnaire were administered. Weight loss was defined as percent loss in body weight from highest lifetime weight to measured weight at baseline. RESULTS: Weight loss was associated with older age, higher rates of diseases such as diabetes, and lower baseline levels of blood pressure and serum total cholesterol. After controlling for age, medical, and lifestyle factors, both men and women had higher mortality rates over a 10+ year period for increasing categories of weight loss (hazard ratio [ 95% CI]: 1.16 [1.06, 1.27] for men and 1.23 [1.13, 1.34] for women). Increased mortality rates with increasing weight loss was shown in stratified analyses of age, body mass index (BMI) at highest weight, smoking, and disease status, but did not always reach statistical significance. Persons on weight loss diets within the year prior to baseline did not have increased mortality with increasing weight loss. CONCLUSION: The strong association between weight loss (likely involuntary) and mortality may be a useful way of estimating overall risks to longevity in populations.PMID: 16029840 [PubMed - in process]

=-=-=-=-=-=-=-=-

T. pct35768@...

__________________________________________________

[Guest guest]

Hi All,

Perhaps providing the below excerpts from the avaialble pdf may be

useful.

Knudtson MD, Klein BE, Klein R, Shankar A.

Associations with weight loss and subsequent mortality risk.

Ann Epidemiol. 2005 Aug;15(7):483-91.

PMID: 16029840

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?

cmd=Retrieve & db=pubmed & dopt=Abstract & list_uids=16029840 & query_hl=2

.... Persons on weight loss diets within the year prior to baseline

did not have increased mortality with increasing weight loss.

CONCLUSION: The strong association between weight loss (likely

involuntary) and mortality may be a useful way of estimating overall

risks to longevity in populations.

.... Analyses stratified by age, BMI at highest weight, disease

status, smoking status, exercise status, and dieting within the year

prior to the baseline examination are presented in Table 3A for men

and Table 3B for women. In general, the trends across these strata

were consistent with men who lost more weight to have greater

mortality rates. However, after multivariate adjustment some trends

were not statistically significant (Table 3A). Men whose BMI at their

highest weight was greater than or equal to 30 kg/m2 only had a

marginally increased death rate with increasing weight loss category

(p = 0.06). Further restricting the analysis to men with BMI > 35

kg/m2 showed a non-significant 13% increase in mortality per

increasing weight loss category (data not shown). Men aged less than

65 years did not have an increased death rate with increasing weight

loss category (p = 0.26). Both non-smoking men and currently smoking

men did not have significantly increased mortality rates with

increasing weight loss (p > 0.20). Similarly, men who reported having

an active lifestyle did not have a statistically significant

increased mortality rate with increasing weight loss (p = 0.21). Men

reporting being on a weight loss diet in the year prior to the

baseline examination did not have a significantly increased mortality

rate with increasing weight loss (HR [95% CI]: 1.02 [0.74, 1.41]).

Further stratification of specific diseases (e.g., diabetes) showed

similar increased mortality rates with increasing weight loss.

However, due to small sample size of specific diseases these hazard

ratios did not reach statistical significance (data not shown).

Table 3A. Stratified analyses investigating weight loss and

mortality in men

.......................................

Strata N Total N Death Mort 10-yr MV-adjusted* HR and 95% CI p-trend

.......................................

Age 43–64 years 1307 174 12.2 1.10 (0.93,1.29) 0.26

Age 65–86 years 779 395 47.9 1.19 (1.06,1.33) 0.002

BMI <25 at highest weight 99 18 16.3 0.79 (0.29,2.10) 0.63

BMI 25–29.9 at highest weight 713 189 25.0 1.38 (1.16,1.64) <0.001

BMI & #8805;30 at highest weight 1274 362 26.6 1.11 (1.00,1.25) 0.06

No known disease † 1275 216 16.1 1.18 (1.03,1.36) 0.02

Has at least one disease † 850 368 40.2 1.17 (1.04,1.31) 0.007

Non-smokers 571 119 19.3 1.01 (0.83,1.23) 0.90

Past smokers 1046 307 27.7 1.29 (1.14,1.46) <0.001

Current smokers 469 143 28.4 1.12 (0.94,1.34) 0.20

Active lifestyle 519 121 21.4 1.14 (0.93,1.40) 0.21

Sedentary lifestyle 1567 448 26.9 1.15 (1.04,1.28) 0.006

Special diet to lose weight ‡ 214 63 27.2 1.02 (0.74,1.41) 0.92

.....................................

N total (death) = total number of persons (deaths).

Mort 10-yr = Kaplan–Meier estimated failure at 10 years (multiplied

by 100).

HR = hazard ratio; CI = confidence interval.

* Multivariate adjusted hazard ratio for a one step increase in

weight loss category adjusted for age, smoking, current drinking

status, cardiovascular disease history, diabetes, BMI at highest

weight, systolic blood pressure, systolic blood pressure2, white

blood cell count, serum creatinine, HCT, proteinuria, exercise, and

reliability of weight.

† Diseases considered include a history of diabetes, gross

proteinuria, non-skin cancer, cardiovascular disease, gout, and/or

thyroid disease.

‡ Only asked amongst persons who reported losing 10 or more pounds

in the year prior to the baseline examination.

Table 3B. Stratified analyses investigating weight loss and mortality

in women

......................................

Strata N Total N Death Mort 10-yr MV-adjusted * HR and 95% CI p-trend

......................................

Age 43–64 years 1469 97 6.2 1.20 (0.97,1.48) 0.09

Age 65–86 years 1210 458 35.1 1.23 (1.13,1.35) <0.001

BMI <25 at highest weight 435 57 11.8 1.79 (1.31,2.44) <0.001

BMI 25–29.9 at highest weight 951 178 16.7 1.35 (1.14,1.59) <0.001

BMI & #8805;30 at highest weight 1263 302 22.7 1.15 (1.04,1.27) 0.009

No known disease† 1575 222 12.7 1.30 (1.14,1.48) <0.001

Has at least one disease† 1166 360 29.4 1.25 (1.13,1.38) <0.001

Non-smokers 1556 342 20.5 1.20 (1.08,1.33) <0.001

Past smokers 647 120 17.6 1.35 (1.13,1.63) 0.001

Current smokers 476 93 17.2 1.24 (1.01,1.51) 0.04

Active lifestyle 609 86 13.4 1.19 (0.97,1.47) 0.10

Sedentary lifestyle 2070 469 21.0 1.24 (1.13,1.35) <0.001

Special diet to lose weight ‡ 256 50 17.7 1.14 (0.81,1.59) 0.45

..............................................

N total (death) = total number of persons (deaths).

Mort 10-yr = Kaplan–Meier estimated failure at 10 years (multiplied

by 100).

HR = hazard ratio; CI = confidence interval.

* Multivariate adjusted hazard ratio for a one step increase in

weight loss category adjusted for age, smoking, cardiovascular

disease history, diabetes, white blood cell count, serum creatinine,

gross proteinuria, serum albumin, less than high school education,

cancer history, serum HDL cholesterol, and reliability of weight.

† Diseases considered include a history of diabetes, gross

proteinuria, non-skin cancer, cardiovascular disease, gout, and/or

thyroid disease.

‡ Only asked amongst persons who reported losing 10 or more pounds

in the year prior to the baseline examination.

Stratified analyses in women showed increasing mortality rates with

increasing weight loss in women aged at least 65 years, any BMI at

highest weight, regardless if a known disease was present, and any

smoking status (Table 3B). Further restricting the analysis to women

with BMI > 35 kg/m2 showed a non-significant 12% increase in

mortality per increasing weight loss category (data not shown). Women

who reported having an active lifestyle also did not have a

statistically significant increased mortality rate with increasing

weight loss (p = 0.10). Women reporting being on a weight loss diet

did not have a significantly increased mortality rate with increasing

weight loss (multivariate HR [95% CI]: 1.14 [0.81, 1.59]). Women aged

less than 65 years had a marginally increased risk of death for each

weight loss category (p = 0.09). As was true in men, further

stratification of specific diseases showed similar increased

mortality with increasing weight loss, but these results did not

reach statistical significance (data not shown).

We investigated the relationship of weight loss and mortality further

in women and men aged at least 65 years. Results were consistent with

those displayed in Tables 2 and Table 3A and Table 3B (data not

shown). In a further attempt to exclude all persons with a

potentially fatal disease at baseline, all analyses described above

were repeated after excluding deaths within the first 2 years of

follow-up. There were no changes to any of the associations. For

example, after excluding early deaths and adjusting for age, smoking,

lifestyle, and other medical factors, for each increase in weight

loss category the risk of death increased by 14% for men (HR and 95%

CI of 1.14 [1.04, 1.26], p = 0.007) and 21% for women (HR and 95% CI

of 1.21 [1.10, 1.32], p < 0.001).

The majority of deaths in the population had a mention of

cardiovascular disease on the death certificate (n = 742, 64% of all

deaths). Analyzing cardiovascular disease mortality resulted in

similar associations of weight loss to all-cause mortality (data not

shown). We also investigated 175 cancer-related deaths. The

association with increasing weight-loss category and cancer mortality

was not significant in men, but was in women (multivariate HR, 95%

CI: 1.07 [0.91, 1.27] for men and 1.21 [1.04, 1.41] for women).

Discussion

For the current report, we investigated associations of various

personal characteristics with weight loss from highest self-reported

weight and subsequent mortality risks in a large cohort of adults

aged between 43 and 86 years. Weight loss was associated with lower

levels of blood pressure and serum total cholesterol. In contrast,

those with a greater weight loss were more likely to have diabetes

and/or a history of cardiovascular disease. Whether these

observations, which are cross-sectional, result from different

antecedent-consequent relationships cannot be determined from our

data.

Persons who lost more of their highest lifetime weight had higher

mortality rates. In general, this increased mortality risk was

consistent in stratified analyses of age, BMI, smoking, and disease

status, but did not always reach statistical significance. Persons

known to be on a special weight loss diet did not have increased

mortality rates for increasing weight loss.

Weight loss has often been used as a marker for disease severity.

This was evident in our data in that persons with either diabetes,

history of cancer, history of cardiovascular disease or other known

diseases had higher mortality rates if they had lost weight than if

they were close to their highest lifetime weight. When we stratified

individually by specific diseases (e.g., diabetes), the association

between weight loss and mortality was not statistically significant.

However, the prevalence of specific diseases was low limiting our

power to detect associations. Consistent with this, several studies,

including the NHANES I population, after adjusting for pre-existing

illness and excluding early deaths found an increased mortality risk

for persons who lost weight 1, 2, 3, 6, 7, 8, 10, 11 and 12.

Studies that measure voluntary versus involuntary weight loss have

generally shown that voluntary weight loss or attempts to voluntarily

lose weight, regardless of the weight loss, have shown decreased

mortality rates, while involuntary weight loss has been consistently

associated with increased mortality rates 10, 14, 17, 24, 25 and 26.

Only a few researchers have found voluntary weight loss to be

associated with increased mortality (8). This may be explained by

uncontrolled confounding since in the Yarri et al. study (8) they

only controlled for age. This may further be explained by the

inability to truly measure voluntary versus involuntary weight loss,

as weight loss may be a composition of both 17 and 27. In our study,

both men and women who reported being on a special diet within one

year of the baseline examination did not have a significant increased

mortality risk after adjusting for age, medical, and lifestyle

factors. This suggests most of the realized weight loss in our study

reflects other processes (e.g., severity of disease) associated with

mortality. The reason for not finding a decrease in mortality in

persons reporting being on weight loss diets may be a result of our

crude measurement of voluntary weight loss.

Several studies have shown that the weight loss and mortality risk is

dependent upon initial BMI 2, 9 and 13. A study by Somes et al. (13)

showed decreased mortality risks in obese persons who lose weight. In

contrast, we showed increased mortality rates for increasing weight

loss for both obese and overweight persons. The long-term effects of

body mass index and obesity have been well studied and generally show

obesity to be associated with adverse outcomes (28). However, a

single measure of body mass at a baseline examination may not be

sufficient to study the effects of weight on mortality. For example,

obese persons who lost weight prior to a baseline examination may

have a low BMI, as was evident in our analysis (see Table 1). This

may explain, in part, why some studies show a U-shaped relationship

with BMI and mortality (5). Another consideration is that we did not

measure body composition to distinguish between fat and fat-free

mass. Recent research has shown that fat mass increases mortality

risk and fat-free mass decreases mortality risk 17 and 29. These

measurements are inherently combined in the BMI and may be an

explanation why different mortality risks with weight loss were not

shown stratifying by BMI.

Overall mortality findings from our study seem to reflect the risk of

cardiovascular disease death as is true for other studies

investigating cause-specific mortality 2, 3, 4, 8 and 12. We did not

find an association between weight loss and cancer-related mortality

in men. Yaari et al. (8) also did not find a relationship in cancer-

mortality in elderly men. Most cancers are strongly linked to

smoking. The excess risk attributable to weight loss may have been

lower due to these other strong factors.

Strengths of this study include the large size of the population

studied and the ability to investigate stratified analyses of many

potentially confounding factors. Investigation of these smaller

groups aids in giving a better understanding of the associations,

rather than adjustment in multivariate models. However, our study has

several limitations. The population studied is relatively homogeneous

in nature (99% white) and confined to a small geographic area of the

United States and, therefore, may not be generalizable to populations

of a different race or ethnicity. Another limitation is that we only

have information on highest weight. We do not know if the weight loss

was recent or if a significant weight gain occurred prior to the

baseline examination, nor do we know if a person's weight had cycled.

This may explain, in part, why excluding deaths within the first few

years of follow-up did not alter the associations. Other researchers

have shown similar associations after excluding deaths in early

follow-up 3 and 4. Results may have been stronger if we had

information on weight cycling as it has been shown that weight

cycling is related to increased mortality 16, 30 and 31.

Additionally, we cannot distinguish whether or not a decrease of fat-

free or fat mass leads to the increased mortality rates in our study.

We do not know the height of the person at their highest weight. It

has been shown that height decreases with age (32). Therefore, at the

highest weight the person may have been slightly taller and thus a

minor weight loss may have resulted in no change to the BMI. Another

limitation was that highest weight was based on self-report. Sicker

persons may have been more likely to recall being at a higher weight

previously than persons without any illness.

In conclusion, we found an association between weight loss from

highest lifetime weight and increased mortality risk. We could not

fully distinguish between voluntary and involuntary weight loss.

However, persons who reported being on special diets to lose weight

who lost weight were not at increased risk of dying. Future

epidemiological studies investigating mortality should be designed to

inquire about weight loss history including whether weight loss was

secondary to diet. Studies cannot possibly inquire about the presence

and severity of all diseases. ...

> ----- Original Message -----

> From: T

>

> Sent: Sunday, July 24, 2005 3:21 AM

> Subject: [ ] Weight loss, or Weight loss " Diet " ?

Your local actuary wants to know......

>

>

> Hello out there in Beaver Dam!

>

> " After controlling for age, medical, and lifestyle factors, both

men and women had higher mortality rates over a 10+ year period for

increasing categories of weight loss...... "

>

> " Persons on weight loss diets within the year prior to baseline

did not have increased mortality with increasing weight loss.

CONCLUSION: The strong association between weight loss (likely

involuntary) and mortality may be a useful way of estimating overall

risks to longevity in populations. "

>

> I think this type of research is important, if only for

highlighting the limitations of epidemiologic research. You have to

carefully control for this type of thing, or, invariably, you will

get the " frail " thrown in with the the folks innocently following the

Schwarzbein Principle.

>

> =-=-=-=-=-=-=-=-=-=-=--=-=-=

>

> Ann Epidemiol. 2005 Aug;15(7):483-91.

> Associations with weight loss and subsequent mortality risk.

> Knudtson MD, Klein BE, Klein R, Shankar A.

> PMID: 16029840 [PubMed - in process]