Effects of Weight Loss in Overweight/Obese Individuals and Long-Term Hypertension Outcomes
A Systematic Review
Many studies have assessed short-term effects of weight loss on blood pressure, whereas little attention has been paid to long-term effects. We conducted a systematic review to evaluate the long-term effects of weight loss on hypertension outcome measures in adults using literature published from 1966 to 2001. All prospective studies and trials, performed on participants with body mass index of ≥28 kg/m2 with a follow-up of >2 years and weight changes recorded, were included. The data from these studies were used to model the long-term effects on blood pressure. Previous reviews on shorter-term studies indicate a 1:1 drop in blood pressure (mm Hg) with weight loss (kilograms). Our findings, based on studies with follow-up of ≥2 years, demonstrate blood pressure decreases less than this after weight loss. The surgical intervention studies exhibited huge weight losses with undramatic blood pressures changes. When surgical interventions are excluded, the models suggest that for 10 kg weight loss, decreases of 4.6 mm Hg and 6.0 mm Hg in diastolic and systolic blood pressure, respectively, may be expected, about half of that predicted from the short-term trials. Initial blood pressure, the length of follow-up, medication changes, and physiological restrictions may contribute to this reduced effect in the long-term studies. Extrapolation of short-term blood pressure changes with weight loss to the longer term is potentially misleading. The weight/hypertension relationship is complex and needs well-conducted studies with long-term follow-up to examine the effects of weight loss on hypertension outcomes.
Obesity is a chronic health problem affecting increasing numbers of people worldwide and is now recognized as a global epidemic. The United States has had alarming rises in obesity in the last 15 years, with one third of the population affected. Similar obesity trends are seen across Europe.
Many serious medical problems, including hypertension, which predisposes to cardiovascular disease,1 are associated with obesity. In adults, the occurrence of hypertension rises with increasing body weight.2 The Framingham Offspring Study reported that obese women in their 40s were 7× more likely to develop hypertension than their lean counterparts.3 Prevention of obesity would be the ideal solution, but it does not address the risks for those already obese. Most now believe, particularly for the morbidly obese, it is unrealistic to expect weight reduction to a “normal” weight category. Instead, regardless of intervention, clinicians now encourage 5% body weight loss or more as being achievable and maintainable in the long term while resulting in health benefits.
A primary concern in the medical management of obesity is morbidity and mortality risk reduction by improving underlying cardiovascular risk factors such as hypertension and quality of life. Previous intervention studies4 and reviews5 on the effects of intentional weight loss studies have reported that moderate weight losses (losses of 5% to 10%) are associated with significant improvements in obesity-related cardiovascular risk factors such as hypertension and diabetes. More recently, a review of randomized control trials by Neter6 reported a diastolic reduction of 0.92 mm Hg per kilogram of weight loss. However, the primary studies included in these reviews have small sample sizes, have relatively short-time follow-up, or were monitored at the maximum blood pressure reduction often <6 months after the weight loss. There is a lack of evidence of long-term benefits of weight loss on hypertension. The aim of this research was to systematically review the evidence for the long-term effectiveness of weight loss on hypertension in obese populations.
As part of a Health Technology Assessment review, a systematic search of literature was undertaken to evaluate the long-term effects of weight loss in humans on various health outcomes for those with obesity. All prospective and randomized trials from 1966 to 2001 were searched on Medline Embase, Cinahl, and HealthSTAR data bases, details of which have been published.7 This article concentrates on blood pressure outcomes. Reference to obese individuals includes white adults at the top end of overweight as well as to the clinically obese (body mass index >30 kg/m2). Although, follow-up of >5 years was originally specified, the paucity of long-term studies for nonsurgical interventions meant reducing this to 2 years for these studies.
To investigate blood pressure decreases related to weight loss in the long term, weight and blood pressure differences were considered for significance. Differences between follow-up and baseline as paired differences were required but only given by some of the selected articles. The remaining only provided the means of each variable at each time point with a measure of precision. For such cases, the difference between the means at the beginning and end time points acted as a crude estimate of the mean difference. The associated measure of variation for each mean difference was also estimated, using the sum of the variances at the 2 time points as a conservative estimate of variance of difference.
Correlation and linear regression, to investigate and predict differences for diastolic and systolic blood pressure from the associated weight differences, were conducted. Because only mean differences of groups were available (or estimable) rather than differences for all individuals, the relationship between blood pressure and weight changes was better investigated by meta-regression using weighted least squares regression as suggested by Sutton et al.8 Weights were defined as the inverse of the SE of the mean differences of the dependent variable, in this case either the diastolic or systolic differences. The precision of the regression coefficients generated from such a model require further adjustment for meta-regression to determine significance of the coefficients.8
A total of 7567 abstracts were scanned, and 288 full articles were obtained and critically appraised, of which 14 studies met the inclusion criteria and investigated the long-term effects of weight loss on blood pressure. Table 1 shows the basic characteristics of these 14 studies. Eleven studies reported quantitative results, of which 29–10⇓ also had qualitative results. The remaining 3 had only descriptive results of blood pressure changes.11–13⇓⇓ The quantitative blood pressure results were considered by interventions: nonsurgical (Tables 2 and 3⇓) and surgical (Table 4). Nonsurgical interventions are further subdivided as nondrug intervention studies, which include prospective cohort and randomized controlled trials (RCTs; Table 2), and drug intervention studies, which include only RCTs (Table 3).
Of the 7 nonsurgical studies, 214–15⇓ were prospective cohort studies with lifestyle management interventions in which 1 was a workplace with canteen involvement and the other was an intensive residential regimen. Both reported modest but significant reductions in weight and blood pressures.
A nondrug RCT16 retrospectively regrouped its participants according to weight cycling status determined after the intervention at the end of the study (Table 2): (1) gainers gained 4.5 kg from base to 30 months (n=15); (2) the stable group remained within ±4.5 kg of base weight throughout study (n=25); (3) large cyclers lost ≥9 kg initially but returned to ±4.5 kg of base weight at end (n=31); (4) small cyclers lost 4.5 to 9 kg initially and returned to ±4.5 kg of base weight at end (n=28); (5) partial cyclers lost ≥9 kg initially and kept off 4.5 to 9 kg by end (n=28); (6) small successes lost 4.5 to 9 kg initially and kept off 4.5 to 9 kg by end (n=7); (7) and large successes lost ≥9 kg initially and kept off ≥9 kg by end (n=14).
Significant weight losses and blood pressure reductions were seen for partial cyclers (lost ≥9 kg initially and kept off 4.5 to 9 kg). Small successes (lost 4.5 to 9 kg initially and kept off 4.5 to 9 kg) did not have statistically significant reductions (despite the clinically reasonable reductions for weight and blood pressures), probably because of the small sample size (n=7) for this group. Large successes (lost ≥9 kg initially and kept ≥9 kg off) saw a significant weight loss of 12.6 kg; however, the blood pressure reductions for this group were statistically nonsignificant.
Another nondrug RCT17 with a 2-year follow-up used diet and exercise as an intervention. This study showed small nonsignificant decreases in weight and blood pressure measurements.
Three RCTs18–20⇓⇓ investigated drug interventions for obesity (Table 3). These RCTs, along with the nondrug interventions, imply that because more weight is lost, the greater the blood pressure reduction.
Four studies had surgical intervention studies10–9,21–22⇓⇓⇓ (Table 4). All these studies exhibited >20 kg weight loss except 1,22 which had modest weight losses averaging 7.09 kg. However, this latter study included nonmorbidly obese participants without surgical intervention if conservative treatment was successful. It is not clear from the article22 how many participants underwent surgery. This would explain the “small” weight reductions compared with the large amount usually associated with surgical interventions. For these surgical studies, blood pressure reductions were inconsistent and small given the weight losses.
In total, 5 studies9–13⇓⇓⇓⇓ gave qualitative results about hypertension, which are given in Table 5. These qualitative studies had surgical interventions. Consequently, although not specifically given, it may be assumed that the weight losses were large and significant. Hypertension was resolved in 33% to 66% of patients, whereas 22% to 34% remained hypertensive. Others, although remaining on treatment for hypertension, were maintained on lower doses of antihypertensive medication.
Returning to the quantitative measures, the relationship between long-term weight loss and blood pressure changes were investigated. In light of the strongest linear relationship being between weight loss and diastolic blood pressure, the diastolic results are presented first. Figures 1 and 2⇓ show scatter plots of weight differences over the follow-up period, with diastolic and systolic blood pressure differences, respectively. These plots illustrate a different response for the surgical interventions compared with nonsurgical interventions. Despite the dramatic weight losses, the surgical interventions have proportionately smaller blood pressure reductions compared with nonsurgical interventions. With only 3 studies definitely using surgical interventions, the relationship between weight lost by surgical means and blood pressure changes would be poorly defined. Consequently, only weight changes affecting blood pressure for the nonsurgical intervention studies were investigated fully.
Several variables thought to be affecting changes in blood pressure, including weight changes, were considered, correlations of which are given in Table 6. Although the best relationship was between the percentage changes for weight and diastolic blood pressure (r=0.702; P<0.01), the results are similar to those for the absolute differences (r=0.661; P<0.01). Consequently, a regression model was developed using the absolute differences as the more informative to interpret (Table 7). To account for group data and not individuals, the inverse of the SEs of the diastolic differences was used in a weighted least squares regression, giving the model:
Once adjusted for the meta-regression, this model indicates that weight difference has a highly significant role, whereas the constant may be zero. The usual assumptions were upheld.
For comparability, a similar-weighted least squares model was developed on the absolute systolic differences as the dependent variable (weighted according to the inverse of SE for systolic differences of each subgroup; Table 7) and the absolute weight differences as the independent variable:
Although not significant as an ordinary regression model, once adjusted for the meta-regression, the weight difference coefficient is significant. However, the fit of this model is poor (adjusted R2=0.155), and other assumptions also indicate this to be a nonreliable predictive model.
The relationship between blood pressure changes and follow-up times was also investigated (Figures 3 and 4⇓). Diastolic blood pressure changes are related to weight changes (labeled in the plots) up to ≈3 years, after which the relationship breaks down and blood pressure appears to creep up with time. This time effect is also reflected to some extent in the systolic blood pressure results.
As expected, initial blood pressure was highly related to the follow-up blood pressure measurements, as shown in Figures 5 and 6⇓. Thus, someone with high blood pressure initially will still have relatively high blood pressure at follow-up despite reductions.
This review attempts to determine weight loss effects over ≥2 years on hypertension and blood pressure. Of the 14 studies finally selected, 5 had qualitative results relevant to hypertension, whereas 11 of them had quantitative results relating to blood pressure and weight loss. The focus was on quantitative results because these provide data to model the relationship between weight loss and blood pressure change. The diastolic model gave a good fit, satisfying normal assumptions, indicating that a 10-kg weight loss might decrease the diastolic blood pressure by 4.6 mm Hg. Similarly, for systolic blood pressure, 10-kg weight loss might give a 6.0-mm Hg decrease in systolic blood pressure, although this is a less adequate model.
Despite ignoring surgical interventions, both of the estimated models gave smaller predicted reductions of blood pressure than those by the Scottish Intercollegiate Guidelines Network (SIGN)23 and a recent review by Neter,6 both of which indicated a 1:1 relationship for diastolic blood pressure and at least that for systolic blood pressure. However, these were based on studies with follow-up of only 1 year.
Unlike other health outcomes24–25⇓ that benefit from weight reduction, the surgical intervention studies did not show blood pressure reductions proportional to weight loss. There are various reasons for this, such as the possibility that the blood pressure relationship with weight loss for those with extreme weight losses may not be linear, and initial blood pressure could be a confounding factor. Only patients with borderline hypertension may have been selected for surgery, hence blood pressure–lowering effects would not have been as pronounced. Changes in medication administered may also have been a confounding effect; this is indicated by the qualitative results for hypertension.9–13⇓⇓⇓⇓ The less-than-expected effect in the surgical studies could be attributable to the longer follow-up time in these studies, with blood pressure reverting back to the original high blood pressure.7 Certainly, for morbidly obese groups associated with surgical interventions, it would be inadvisable to extrapolate the results to predict blood pressure changes from weight losses given that they were excluded from the modeling process.
The initial average diastolic blood pressure levels were all <95 mm Hg for all subgroups. These averages obviously represent a range of individual values not available for this review. Although the subgroup averages give better point estimates when considered over several studies, this loss of information may be important given the apparent link with initial and follow-up blood pressure.
In general, weight loss results in decreased blood pressure. The maximum benefit is soon after initial weight loss. However, this review has shown that in the long term, the relationship between weight loss and blood pressure change is not consistent with the shorter-term studies or even with the relationship between weight loss and other health outcomes. Clearly, weight/hypertension/blood pressure have a complex relationship with each other and with other variables, including lifestyle factors and medication. A feature that stands out is that blood pressure before weight loss is highly influential, indicating the importance of prevention of hypertension in the first place or indeed the importance of preventing weight gain.
This review was funded by the National Health Service (NHS) R&D Health Technology Assessment Programme. The Health Services Research Unit, University of Aberdeen, is core-funded by the Chief Scientist Office of the Scottish Executive Health Department. We thank Tamara Brown, research fellow, School of Health and Social Care, University of Teesside, Middlesbrough; Professor Adrian Grant, Health Services Research Unit, University of Aberdeen; and Professor Mo Malek (deceased), Health Economics, University of St. Andrews, for their contributions.
- Received November 17, 2004.
- Revision received December 4, 2004.
- Accepted March 31, 2005.
- ↵Neter JE, Stam BE, Kok FJ, Grobbee DE, Geleijnse JM. Influence of weight reduction on blood pressure: a meta-analysis of randomized controlled trials. Hypertension. 2003; 42: 878–884.
- ↵Avenell A, Broom J, Brown TJ, Poobalan A, Aucott L, Stearns SC, Smith WC, Jung RT, Campbell MK, Grant AM. Systematic review of the long-term effects and economic consequences of treatments for obesity and implications for health improvement. Health Technol Assess. 2004; 8: 21.
- ↵Sutton AJ, Abrams KR, Jones DR, Sheldon TA, Song F. Systematic reviews of trials and other studies. Health Technol Assess. 1998; 2: 81–84.
- ↵Sjostrom CD, Peltonen M, Wedel H, Sjostrom L. Differentiated long-term effects of intentional weight loss on diabetes and hypertension. Hypertension. 2000; 36: 20–25.
- ↵O’Leary JP. Overview: jejunoileal bypass in the treatment of morbid obesity. Am J Clin Nutr. 1980; 33: 389–394.
- ↵Sjostrom M, Karlsson AB, Kaati G, Yngve A, Green LW, Bygren LO. A four week residential program for primary health care patients to control obesity and related heart risk factors: effective application of principles of learning and lifestyle change. Eur J Clin Nutr. 1999; 53 (suppl 2): S72–S77.
- ↵Wing RR, Venditti E, Jakicic JM, Polley BA, Lang W. Lifestyle intervention in overweight individuals with a family history of diabetes. Diabetes Care. 1998; 21: 350–359.
- ↵Scottish Intercollegiate Guidelines Network. Obesity in Scotland. Integrating prevention with weight management. A national clinical guideline recommended for use in Scotland. Edinburgh, Scotland: SIGN; 1996; 8.