Direction of Weight Change in Recurrent Depression Consistency Across Episodes Albert J.
Stunkard, MD; Madelyn H. Fernstrom, PhD; R. Arlen Price, PhD; Ellen Frank, PhD; David J. Kupfer MD
\s=b\ The direction and extent of weight change during two separate episodes of severe, unipolar depression were assessed in 53 (unmedicated) outpatients in the Pittsburgh (Pa) study of maintenance therapy of depression. There was a high concordance (45 of 53 patients) of direction of self-reported weight change during the two episodes. Twenty-three patients lost weight during both episodes, 17 gained weight, and 5 showed no change. The extent of weight change between the two episodes was highly correlated. Self-reported weight change corresponded closely to measured weight changes in a large sample of the study population. Changes in appetite paralleled those in body weight. Duration of the episode and body mass index were related to the weight change, but two features of depression with which weight loss in depression has been associated (the endogenous character of the depression and its severity) were not. Direction and extent of weight change in unipolar depression appear to be stable patient characteristics across episodes and are thus potential markers for subtypes of depression. This stability of weight change is in sharp contrast to the lack of stability of the endogenous subtype in consecutive episodes of major depression. (Arch Gen Psychiatry. 1990;47:857-860)
many as 40% of depressed patients (and weight loss in as few as 30%). Weight gain has been ascribed to two sources— lack of severity of the depression and two possibly related types" of depression. One, with a distinctive clinical profile, the dieting depression," afflicts primarily obese women.16 The other is "atypical depression," characterized by hypersomnia, a low level of energy, and an early age at onset.11"'23,24 The fact that weight may be either lost or gained during depression raises the possibility that the direction, and per¬ haps the extent, of weight change may help to differentiate types of depression. Such a distinction could be useful by providing a more sensitive measure of depression than the current use of weight loss only, which is still used in depres¬ sion inventories.2'7,8 If weight change in depression is to serve these functions, it must be a stable patient characteristic: persons must consis¬ tently either lose or gain weight during different episodes of depression. Surprisingly, it is not known whether the direc¬ tion of weight change in depression is a stable characteristic. The present study was undertaken to find that out.
Weight
Data were obtained from the Pittsburgh (Pa) Maintenance Thera¬ pies in Recurrent Depression, a project that enrolled unipolar pa¬ tients with a high probability of recurrence.26 Patients were between 21 and 65 years of age and in their third or later episode of unipolar major depression, with a minimum Hamilton Depression Rating Scale8 score of 15 and a minimum Raskin26 score of 7. All patients received the same acute treatment regimen, which consisted of imip¬ ramine hydrochloride (150 to 300 mg/d) and interpersonal psycho¬ therapy. After treatment they were assigned to one of five mainte¬ nance treatments, conducted on an outpatient basis: a maintenance form of interpersonal psychotherapy offered alone, maintenance in¬ terpersonal psychotherapy with imipramine continued at the acute treatment dose, maintenance interpersonal psychotherapy with pla¬ cebo, medication clinic with imipramine, or medication clinic with placebo. The present study enrolled 68 patients in two subgroups. For one subgroup, ("new patients," 37), the index episode of depression was that which initiated their entry into the program and the "previ¬ ous episode" was the one just before it. For the other subgroup ("continuing patients," 31), the index episode was a recurrence after they had entered the program and the "previous episode" was
loss has been recognized as a sign of depression1'10 since Kraepelin's description in 1904,3 and two repre¬ sentative reports have described weight loss in 78% and 80% of depressed patients.910 Loss of weight has been ascribed to two aspects of depression, severity2,7,8 and endogenous char¬ acter of the depression.11"13 There have been, however, no studies of severity as a predictor of weight loss in depression, and Nelson and Charney's14 review of 13 studies showed that endogenous character was only weakly associated with weight loss. More recently, weight gain has been reported in depres¬
sion,15"21 with Weissenburger et al22 describing weight gain in
Accepted for publication January 10,1990. From the Departments of Psychiatry, University of Pennsylvania, Philadelphia (Drs Stunkard and Price) and University of Pittsburgh (Pa) (Drs Fernstrom, Frank, and Kupfer). Reprint requests to Department of Psychiatry, University of Pennsylvania School of Medicine, 133 S 36th St, Suite 507, Philadelphia, PA 19104-3246 (Dr Stunkard).
as
PATIENTS AND METHODS
Downloaded From: http://archpsyc.jamanetwork.com/ by a Cornell University User on 08/01/2016
=
=
Table 1. —Characteristics of the
lated the average weight during the 6 months before the recurrence for each patient and compared it with the weight measured at the time
Sample Mean ± SD ±11.6
Age, y_67_45.1 Height, m Men
12
Women
56
1.76 + 0.08 1.64 ±0.06
Weight, kg
Men_12_80.7 ±11.9 Women_56_70.7 ±13.5
Body mass index Weight change, kg (current episode) Absolute weight change, kg (current episode) Weight change, kg (previous episode)
68
26.3 ±4.0 0.57 ± 6.2
67 -
67
5.1 ±3.6
44
0.39 ± 6.8 -
Duration,
mo
Current episode
episode Mood, current episode Appetite, current episode Previous
Table 2. —Weight
Episode
5.0 ±5.8
19
6.0 ±6.2
67
8.9 ±1.0
65
3.8 ±3.4
Changes During the Index and Previous Episodes* Index
Previous
57
Episode
'-
Loss
No
Change
Gain
Total
Loss_23_2_1_26 No
change
0
5
0
5
Gain_4_1_17_22_
Total
»Likelihood ratio
27
8
18
53
2(4) 54.55, P
Stunkard, MD; Madelyn H. Fernstrom, PhD; R. Arlen Price, PhD; Ellen Frank, PhD; David J. Kupfer MD
\s=b\ The direction and extent of weight change during two separate episodes of severe, unipolar depression were assessed in 53 (unmedicated) outpatients in the Pittsburgh (Pa) study of maintenance therapy of depression. There was a high concordance (45 of 53 patients) of direction of self-reported weight change during the two episodes. Twenty-three patients lost weight during both episodes, 17 gained weight, and 5 showed no change. The extent of weight change between the two episodes was highly correlated. Self-reported weight change corresponded closely to measured weight changes in a large sample of the study population. Changes in appetite paralleled those in body weight. Duration of the episode and body mass index were related to the weight change, but two features of depression with which weight loss in depression has been associated (the endogenous character of the depression and its severity) were not. Direction and extent of weight change in unipolar depression appear to be stable patient characteristics across episodes and are thus potential markers for subtypes of depression. This stability of weight change is in sharp contrast to the lack of stability of the endogenous subtype in consecutive episodes of major depression. (Arch Gen Psychiatry. 1990;47:857-860)
many as 40% of depressed patients (and weight loss in as few as 30%). Weight gain has been ascribed to two sources— lack of severity of the depression and two possibly related types" of depression. One, with a distinctive clinical profile, the dieting depression," afflicts primarily obese women.16 The other is "atypical depression," characterized by hypersomnia, a low level of energy, and an early age at onset.11"'23,24 The fact that weight may be either lost or gained during depression raises the possibility that the direction, and per¬ haps the extent, of weight change may help to differentiate types of depression. Such a distinction could be useful by providing a more sensitive measure of depression than the current use of weight loss only, which is still used in depres¬ sion inventories.2'7,8 If weight change in depression is to serve these functions, it must be a stable patient characteristic: persons must consis¬ tently either lose or gain weight during different episodes of depression. Surprisingly, it is not known whether the direc¬ tion of weight change in depression is a stable characteristic. The present study was undertaken to find that out.
Weight
Data were obtained from the Pittsburgh (Pa) Maintenance Thera¬ pies in Recurrent Depression, a project that enrolled unipolar pa¬ tients with a high probability of recurrence.26 Patients were between 21 and 65 years of age and in their third or later episode of unipolar major depression, with a minimum Hamilton Depression Rating Scale8 score of 15 and a minimum Raskin26 score of 7. All patients received the same acute treatment regimen, which consisted of imip¬ ramine hydrochloride (150 to 300 mg/d) and interpersonal psycho¬ therapy. After treatment they were assigned to one of five mainte¬ nance treatments, conducted on an outpatient basis: a maintenance form of interpersonal psychotherapy offered alone, maintenance in¬ terpersonal psychotherapy with imipramine continued at the acute treatment dose, maintenance interpersonal psychotherapy with pla¬ cebo, medication clinic with imipramine, or medication clinic with placebo. The present study enrolled 68 patients in two subgroups. For one subgroup, ("new patients," 37), the index episode of depression was that which initiated their entry into the program and the "previ¬ ous episode" was the one just before it. For the other subgroup ("continuing patients," 31), the index episode was a recurrence after they had entered the program and the "previous episode" was
loss has been recognized as a sign of depression1'10 since Kraepelin's description in 1904,3 and two repre¬ sentative reports have described weight loss in 78% and 80% of depressed patients.910 Loss of weight has been ascribed to two aspects of depression, severity2,7,8 and endogenous char¬ acter of the depression.11"13 There have been, however, no studies of severity as a predictor of weight loss in depression, and Nelson and Charney's14 review of 13 studies showed that endogenous character was only weakly associated with weight loss. More recently, weight gain has been reported in depres¬
sion,15"21 with Weissenburger et al22 describing weight gain in
Accepted for publication January 10,1990. From the Departments of Psychiatry, University of Pennsylvania, Philadelphia (Drs Stunkard and Price) and University of Pittsburgh (Pa) (Drs Fernstrom, Frank, and Kupfer). Reprint requests to Department of Psychiatry, University of Pennsylvania School of Medicine, 133 S 36th St, Suite 507, Philadelphia, PA 19104-3246 (Dr Stunkard).
as
PATIENTS AND METHODS
Downloaded From: http://archpsyc.jamanetwork.com/ by a Cornell University User on 08/01/2016
=
=
Table 1. —Characteristics of the
lated the average weight during the 6 months before the recurrence for each patient and compared it with the weight measured at the time
Sample Mean ± SD ±11.6
Age, y_67_45.1 Height, m Men
12
Women
56
1.76 + 0.08 1.64 ±0.06
Weight, kg
Men_12_80.7 ±11.9 Women_56_70.7 ±13.5
Body mass index Weight change, kg (current episode) Absolute weight change, kg (current episode) Weight change, kg (previous episode)
68
26.3 ±4.0 0.57 ± 6.2
67 -
67
5.1 ±3.6
44
0.39 ± 6.8 -
Duration,
mo
Current episode
episode Mood, current episode Appetite, current episode Previous
Table 2. —Weight
Episode
5.0 ±5.8
19
6.0 ±6.2
67
8.9 ±1.0
65
3.8 ±3.4
Changes During the Index and Previous Episodes* Index
Previous
57
Episode
'-
Loss
No
Change
Gain
Total
Loss_23_2_1_26 No
change
0
5
0
5
Gain_4_1_17_22_
Total
»Likelihood ratio
27
8
18
53
2(4) 54.55, P
