A substantive amendment to this systematic review was last made on 11 November 1998. Cochrane reviews are regularly checked and updated if necessary.INTRODUCTIONBackground and objectives: To determine whether antidepressants are clinically effective and acceptable for the treatment of depression in people who also have a physical illness.
Search strategy: Medline, Cochrane Library Trials Register and Cochrane Depression and Neurosis Group Trials Register were all systematically searched, supplemented by hand searches of two journals and reference searching.
Selection criteria: All relevant randomised trials comparing any antidepressant drug (as defined in the British National Formulary) with placebo or no treatment, in patients of either sex over 16, who have been diagnosed as depressed by any criterion, and have a specified physical disorder (for example cancer, myocardial infarction). ""Functional"" disorders where there is no generally agreed physical pathology (e.g. irritable bowel syndrome) were excluded. The main outcome measures are numbers of individuals who recover/improve at the end of the trial and, as a proxy for treatment acceptability, numbers who complete treatment.
Data collection and analysis: Data was extracted independently by the reviewers onto data collection forms and differences settled by discussion.
Main results: 18 studies were included, covering 838 patients with a range of physical diseases (cancer 2, diabetes 1, head injury 1, heart 1, HIV 5, lung 1, multiple sclerosis 1, renal 1, stroke 3, mixed 2). Depression was diagnosed clinically in 3 studies, otherwise by structured interview or checklist.
Only 5 studies described how they performed randomisation. 1 study compared drug with no treatment, and the rest with placebo: all of the latter said they were double blind.
6 studies used SSRIs, 3 atypical antidepressants, and the remainder tricyclics.
Patients treated with antidepressants were significantly more likely to improve than those given placebo (13 studies, OR 0.37, 95% CI 0.27-0.51) or no treatment (1 study, OR 3.45, 95% CI 11.1-1.10). About 4 patients would need to be treated with antidepressants to produce one recovery from depression which would not have occurred had they been given placebo (NNT 4.2, 95% CI 3.2-6.4).
Most antidepressants (tricyclics and SSRIs together, 15 trials ) produced a small but significant increase in dropout (OR 1.66, 95% CI 1.14-2.40. NNH 9.8, 95% CI 5.4-42.9). The ""atypical"" antidepressant mianserinproduced significantly less dropout than placebo.
Only 2 studies used numerical scales designed to measure effects on function and quality of life; in HIV (Karnofsky scale), drug was better than no treatment; in lung disease (Sickness Impact Profile), drug was not significantly different from placebo.
Only 7 studies reported looking for changes in the physical disease. Antidepressants produced no change in immune function in HIV relative to placebo (2 studies) or no treatment (1 study). Relative to placebo, antidepressants produced no change in cardiovascular function in heart disease, in respiratory function in lung disease, or in vital signs or laboratory tests in cancer (1 study each). Nortriptyline produced worse control in diabetes.
Trends towards tricyclics being more effective than SSRIs, but also more likely to produce dropout were noted, but these are based on non-randomised comparisons between trials.
Reviewers conclusions: The review provides evidence that antidepressants, significantly more frequently than either placebo or no treament, cause improvement in depression in patients with a wide range of physical diseases.
About 4 patients would need to be treated with antidepressants to produce one recovery from depression which would not have occurred had they been given placebo (NNT 4.2, 95% CI 3.2-6.4).
Antidepressants seem reasonably acceptable to patients, in that about 10 patients would need to be treated with antidepressants to produce one dropout from treatment which would not have occurred had they been given placebo (NNH 9.8, 95% CI 5.4-42.9).
The evidence is consistent across the trials, apart from 2 trials in cancer, where the ""atypical"" antidepressant mianserin produced significantly less dropout than placebo.
Trends towards tricyclics being more effective than SSRIs, but also more likely to produce dropout were noted, but these are based on non-randomised comparisons between trials.
Problems with the evidence include most of the trials use of observers, rather than patients, to decide on improvement, and concentration mainly on symptoms rather than function and quality of life. There is also a possibility of undetected negative trials.
Nevertheless, the review provides evidence that use of antidepressants should at least be considered in those with both physical illness and depression. Regarding diagnosis, the existence of a cheap and readily available treatment for depression should encourage detailed assessment of peristent low mood in the physically ill.
Depressive disorders are commoner in those with physical illness than those without and severe forms are often seen (1). The depression frequently has a poor outcome and patients often have an increased morbidity (2) and mortality (3). Depressed physically ill patients consume more investigations and treatments than the non-depressed, have worse function (4), and are also less likely to comply with treatment (5).
DIAGNOSTIC ISSUES
Nevertheless, there is continuing debate about the exact status of depression of mood and associated symptoms in the physically ill, and hence uncertainty as to whether such patients are ""really depressed"".
The issues surrounding the diagnosis of depression in the physically ill are complex. For example, it is difficult to be sure whether symptoms such as low mood or lack of energy are due to a mental disorder (for example a depressive illness), or to an understandable and expected reaction to the physical illness, or are due to the physical illness itself.
Furthermore, depression itself, even in patients without physical illness, is associated with ""functional"" physical symptoms such as pain, gastrointestinal disturbance, headache, etc..
A comprehensive discussion of these diagnostic issues is beyond the scope of the present review; readers are referred to specialist texts (e.g. 6) for a more detailed account.
CLINICAL ISSUES
In clinical practice, however, it is rarely possible to be certain about the exact status of depressed mood in an individual patient with physical illness. In the face of this uncertainty, the clinical question often resolves itself into: whether a physically ill patient with persistent depression of mood, and often with some other symptoms consistent with depressive illness, should receive a trial of treatment for depression.
Bearing in mind the diagnostic and clinical issues outlined above, it is not surprising that there is evidence that underdiagnosis and undertreatment of depressive disorder in the physically ill often occurs in clinical practice (e.g. 7) . ""Clinicians often withhold potentially effective antidepressant drugs because they assume that the medical illness caused the depression, or they consider the medical illness to be a contraindication to antidepressant drug treatment"" (8).
Clinicians are certainly less likely to diagnose depression in the physically ill if they do not believe that an effective and safe treatment is available to them. In fact, there is continuing uncertainty about optimum management of depression in the physically ill. This is partly because most treatment trials of depression exclude the physically ill.
ISSUES CONCERNING ANTIDEPRESSANTS
Generally speaking, the treatment for depression which is most readily available to clinicians is antidepressant medication. Therefore it is important to examine the evidence as to whether or not antidepressants are effective and safe in the depressed physically ill.
Short of new large randomised controlled trials, it is accepted that a systematic review of all relevant existing placebo-controlled randomised trials provides the least biased estimate of treatment effect. We have therefore attempted to identify all placebo-controlled randomised controlled trials of antidepressants for depression in the physically ill, and review them systematically, according to Cochrane Collaboration methodology.
ISSUES ABOUT STATISTICAL COMBINATION OF THE TRIALS
In this first version of the review, we consider that our search for all extant trials is perhaps our most important objective, as the trials are likely to be widely scattered, and hard to access for many. Putting before patients and clinicians a list, as exhaustive as possible, of relevant trials will therefore, we hope, be of service.
In addition, we have also attempted a statistical combination of the results of the trials
For the purposes of this combination (meta-analysis) we have considered all types of physical illness together, and all antidepressants together. We recognise that some readers may doubt the validity of this ""broad-brush"" approach. However, there is some evidence to support it.
First, the combination is subject to the trials not showing evidence of statistical heterogeneity. If significant heterogeneity is present, it may not be valid to attempt a statistical combination of the results.
Second, regarding the antidepressants, their similarities in clinical effectiveness and in adverse effects are much more striking than their differences (e.g. 9). Furthermore, such combining of antidepressant classes has been done in other published reviews (e.g. 10).
Third, regarding the physical diseases, depressive illnesses are recognised as occurring across the whole range of physical diseases. Although symptoms may vary somewhat between different diseases (see reference 6 for review), it has not been suggested in international classifications (DSM-IV, ICD 10) that diagnostic criteria for depression (or indeed for any mental disorder) should be different in different physical diseases.
We recognise that the process of met-analysis may tend to mask some real differences which may exist between drugs, or between depression in different diseases, and also suspect that some of these possible differences may be important. These are described in the secondary objectives, and will be addressed in later versions of the review.
REFERENCES
1. Katon W, Sullivan M. Depression and chronic medical illness. J Clin Psychiatry 1990;56(supplement):3-11.
2. Lustman PJ, Griffith LS, Clouse RE. Depression in adults with diabetes. Diabetes Care 1988;11:605-12.
3. Frasure-Smith N, Lesperance F, Taljic M. Depression and 18-month prognosis after myocardial infarction. Circulation 1995;91(4):999-1005.
4. Wells KB, Rogers W, Burnam MA. Course of depression in patients with hypertension, myocardial infarction or insulin-dependent diabetes. Am J Psychiatry 1993;150:632-8.
5. Guiry E, Conroy RM, Hickey N, Mulcahy R. Psychological response to an acute coronary event and its effect on subsequent rehabilitation and lifestyle change. Clin Cardiol 1987;10(4):256-60.
6 Rodin G, Craven C, Littlefield C ( 1991 ): Depression in the medically ill. New York: Brunner Mazel.
7 Perez-Stable EJ,Miranda J, Munoz R, Ying Y: Depression in medical outpatients: underrecognition and misdiagnosis Arch.Int.Med. ( 1990 ) 150, 1083-1088.
8 Roose SP, Dalack GW Treating the depressed patient with
cardiovasculr problems J.Clin.Psychiat. ( 1992 ) 53 ( suppl. ) 25-31
9 Song F et al. Selective serotonin reuptake inhibitors: meta-analysis of efficacy and acceptability. BMJ 306, 683-687, 1993.
10. Lima MS,Moncrieff J A comparison of drugs versus placebo for the treatment of dysthymia: a systematic review. Cochrane Library, 1998, Issue 3.
In this first version of the review, we have considered placebo controlled trials of all antidepressants together, and all types of physical illness together.
The secondary objectives, which will be explored in later versions of the review, are as follows:
1) to compare results of treating depression in conditions primarily affecting the brain ( for example stroke and Parkinson s disease ) with conditions not primarily affecting the brain, as the central nervous system pathology may have a direct effect on areas of the brain which influence mood.
2) to compare different types of antidepressants, notably tricyclic antidepressants (TCA s) and selective serotonin re-uptake inhibitors (SSRI s).
3) to compare patients defined as depressed using a self-rating scale, such as the Hospital Anxiety and Depression Scale, with those defined as depressed using a clinical interview such as the Present State Examination.
Trials involved participants of either sex over the age of 16, who have been diagnosed as depressed by any criterion, and have a specified physical disorder ( for example myocardial infarction ).
Exclusion criteria
1) Studies of depression co-existing with disorders characterised by physical symptoms generally agreed to be unaccounted for by physical pathology ( so called ""functional"" or ""psychosomatic disorders"" such as tension headache, irritable bowel syndrome, etc ) will be excluded.
2) Studies of heterogeneous groups such as ""medical inpatients"", ""geriatric rehabilitation patients"" will be excluded unless the authors specifically excluded those patients who were in the group for psychological/ social /psychosomatic conditions or problems.
3) Apart from anxiety and other neurotic conditions, which frequently co-exist with depression, other psychiatric disorders, notably dementia, psychosis and addictive behaviours, will be exclusions.
4) Antidepressants are frequently used for other indications such as analgesia. Trials of antidepressants for pain relief and other indications will be excluded unless their subjects had definite physical pathology and were explicitly diagnosed and treated for depression.
Any antidepressant medication listed in the British National Formulary number 35 March 1998, section 4.3 (tricyclics, heterocyclics, selective serotonin reuptake inhibitors and mono-amine oxidase inhibitors: see Search Strategy for list) compared to placebo or no intervention.
Trials where there was equivalent supportive or other psychotherapy in the two arms will be included (that is, antidepressant plus pscyhotherapy compared with placebo plus pscyhotherapy.
Exclusion criteria
Trials of euphoriants ( e.g. amphetamines ) and adjuvants ( e.g. lithium,
methionine ) will be excluded, as will combination treatements ( e.g ""Motipres"" ) and studies of other drug classes not generally regarded as primarily antidepressants ( e.g. major and minor tranquilisers ).
Dichotomous: The primary outcome will be the number of individuals who had recovered or improved by group. By “recovered”, we mean that the patient fell below the cutoff score for caseness on any diagnostic instrument used, or that they rated themselves or were rated by the investigators as having recovered. “Improved” means any improvement short of recovery.
Continuous outcomes ( e.g. change in depression scale ( endpoint- baseline ) averaged for each group, endpoint score on depression scales averaged for each group ) are frequently reported in antidepressant trials. We will analyse this data in subsequent versions of the review.
We have preferred to start with the dichotomous data, since there are certain difficulties in meta-analysis of continuous data:
(i) it is not at present possible using Cochrane Collaboration software to do a conservative intention to treat analysis by assigning adverse outcomes to non-completers, so such analyses are likely to be more biassed than than those of dichotomous data.
(ii) some studies report means but not measures of dispersal such as a standard deviation.
(iii) Differences in means of groups of patients (continuous) may be less meaningful to users of the review such as individual patients or their advisers, than the chances of significant clinical improvement (dichotomous).
2) Physical illness changes
Deaths will be summarised, and measures of function /quality of life which are applicable across different disases (""generic"" scales, e.g.SIP, SF36, Karnofsky) will be included in a meta-analysis.
Changes specific to each disease will be summarised in text form under ""physical illness changes"" unless they are used in at least two trials.
3) Acceptability of the treatment
a) acceptability of the intervention to the participants, as assessed by numbers of patients completing the trial.
See: Collaborative Review Group search strategy
1. Handsearching of journals ( J Psychosom Res, Gen Hosp Psych to end 1997) by the authors.2 Search of 23/2/1998 Cochrane Depression & Nuerosis Trial Register using the following search strategy: (Keyword) = DEPRESSION* OR DEPRESSIVE-DISORDER* and (Intervention)
= PHARMACOTHERAPY* combined with COMORBID* OR (MEDICALLY ILL) OR (PHYSICALLY ILL).
3) search of Cochrane Library Trial Register 25/6/1998, (CD-ROM, issue 2 1998) using the following search strategy: (antidepress* and depress*).
4) The British National Formulary number 35 March 1998, section 4.3 lists antidepressants currently available. Their generic and proprietary names (below) were used in electronic searches as indicated.
amitriptyline or lentizol or tryptizol or triptafen or amoxapine or asendis or clomipramine or anafranil or dothiepin or perothiaden or doxepin or sinequan or impramine or tofranil or lofepramine or gamanil or nortriptyline or allegron or protriptyline or concordin or trimipramine or surmontil or maprotiline or ludiomil or mianserin or mirtazapine or zispin or trazodone or molipaxin or viloxazine or vivalan or phenelzine or nardil or isocarboxazid or tranylcypromine or parnate or moclobemide or manerix or fluoxetine or prozac or citalopram or cipramil or fluvoxamine or faverin or paroxetine or seroxat or sertraline or lustral or nefazadone or dutonin or venlafaxine or efexor or flupenthixol or depixol or fluanxol or reboxetine or edronax or tryptophan or optimax.
5) Medline Express on Silver Platter was searched 29/6/1998 using the following search term for randomised controlled trials, combined with the term for antidepressants in 4) above
#1 explode RANDOMIZED- CONTROLLED-TRIALS / all subheadings IN MESH
#2 RANDOMIZED CONTROLLED TRIAL IN PTexplode
#3 RANDOM-ALLOCATION / ALL SUBHEADINGSexplode
#4 DOUBLE-BLIND-METHOD / all subheadings in MESHexplode
#5 SINGLE-BLIND-METHOD / ALL SUBHEADINGSexplode
#6 CLINICAL-TRIALS / ALL SUBHEADINGS
#7 CLINICAL TRIAL IN PT
#8 CONTROLLED CLINICAL TRIAL IN PT
#9 RANDOMI*
#10 RANDOM* near2 (ASSIGN* or ALLOCATE*)
#11 CLIN* near TRIAL*
#12 (SINGL* or DOUBL* or TREBL* or TRIPL*) near (BLIND* or MASK*)
#3 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12
Each electronic report was scrutinised and relevant studies retrieved.
6) Search of 23/2/1998 Cochrane Depression & Neurosis Trial Register using the following search strategy: names of antidepressants in 4) above.
7) Reference lists of retreived studies and reviews have been searched.
8) A number of experts in consultation -liaison psychiatry have been consulted.
9. The following specialist book has been handsearched:
(i) Rodin G, Craven C, Littlefield C ( 1991 ): Depression in the medically ill. New York: Brunner Mazel.
Reference management was in Procite for Windows.
STUDY SELECTION Selection criteria were applied liberally in deciding which articles to retrieve from the results of the search strategy. The full-text articles were examined by each reviewer independently to see whether the criteria have been met.ASSESSMENT OF METHODOLOGICAL QUALITY was done according to the Cochrane Collaboration Handbook. Only trials stated to be randomized ( category A or B of the handbook ) were included.
DATA COLLECTION A standard form designed by the authors was used to extract data from the studies. The data from this form was entered into Revman statistical software. Missing information was obtained from investigators whenever possible.
ANALYSES We envisage that the participants, interventions and outcomes in each of the included trials will be sufficiently similar to make combining them clinically meaningful ( see Backgroound ). We therefore intend to use techniques of meta-analysis. As individual trials identified in this field tend to be small, we intend to combine
1) all physical illnesses,
2) all depression diagnoses,
3) all antidepressants and
4) all depression endpoints.
This increases statistical power, and estimates treatment effect across various settings, but involves assuming that the above “lumpings together” are valid.
HETEROGENEITY A test for heterogeneity between the studies will be done; if there is significant between study variation, a random effects approach will be used, and caution urged in interpreting the overall estimate. However, we shall also use a fixed effects model, as in practice there may be little difference in the final result.
CHOICE OF SUMMARY STATISTIC For binary outcome data, we calculated the odds ratio for each trial, and then a pooled odds ratio across appropriate groups of trials ( using Mantel- Haenszel methods ).
They covered a range of physical illnesses:
cancer 2
multiple sclerosis 1
heart disease 1
stroke 3
HIV/AIDS 5
renal failure 1
diabletes mellitus 1
chronic obstructive airways disease 1
mixed diagnoses 1
mixed diagnoses, elderly 1
head injury 1
Depression was diagnosed clinically in 3 studies, otherwise by structured interview or checklist.
The following antidepressants were used(some studies had more than one drug arm):
atypical 3
TCAs 11
SSRIs 6
Only 5 studies described how they performed randomisation.
All of the 17 placebo-controlled studies said they were double blind, but none reported checking on this. The nature of the placebos were not indicated, so they were presumably inert.
The studies were generally short term, with a mean length of treatment of 6.5 weeks ( range 4-12).
Comparison treatments were placebo in all cases, except one, where there was no treatment indicated (concomitant psychotherapy in both arms).
In one study, there was overt psychotherapy in both arms.
Information for subsequent cost-effectiveness analyses: no information on treatment costs was available from the trials.
All of the 17 placebo-controlled studies said they were double blind, buit none reported checking on this. The nature of the placebos were not indicated, so they were presumably inert.
IMPROVEMENT IN DEPRESSION
Combined results
We have expressd the results in terms of lack of improvement. The 13 trials which reported dichotomous outcome data included 691 patients. Of these 366 took an antidepressant and 325 a placebo. Of those 366 who received an antidepressant 177 (48%) failed to improve compared to 229 who failed to respond out of the 325 (70%) patients who received placebo treatment (Odds ratio(OR) 0.37 95% CI 0.27-0.51).
Put another way, about 4 patients would need to be treated with antidepressants to produce one recovery from depression which would not have occurred had they been given placebo (NNT 4.2, 95% CI 3.2-6.4).
Results by method of reporting outcome
Seven trials used a CGI rating of 1 or 2 as a measure of treatment response, that is a moderate or marked improvement in symptoms. Combining these trials gives a very similar OR for lack of response of 0.35 95% CI 0.23-0.53.
Four trials used a >50% fall in initial ratings on the Hamilton Rating Scale for depression. Combining these for lack of response gives an OR of 0.28, 95% CI 0.17-0.46.
One trial used a >50% fall in initial rating on the Beck Depression Inventory. It had an OR for no improvement of 0.54, 95% CI 0.14-2.07.
Lack of improvement in studies by antidepressant type (TCAs & SSRIs).
Eight studies randomised individuals to tricyclics and placebo. Of the 180 people randomised to a tricyclic 88 (49%) failed to improve compared to 123 of the 163 people (75%) taking a placebo who failed to improve (OR 0.34 95% CI 0.22 to 0.54).
Four studies randomised 122 people to an SSRI and 123 to a placebo. 66 of the 122 (54%) on an SSRI failed to improve compared to 78 of the 123 (63%) on placebo (OR 0.67 95% CI 0.4 to 1.13).
LEAVING THE STUDY PRIOR TO COMPLETION
Information on failing to complete the study was available from 17 trials. These studies included 410 patients randomised to antidepressants and 378 randomised to placebo.
Of those taking antidepressants 123 (30%) left the study before completion compared to 101 (27%) of those randomised to placebo; this is not a statistically significant difference (OR 1.15 (95% CI 0.82 to 1.6)). However, there was significant statistical heterogeneity between the studies so the above combined estimate must be interpreted cautiously.
The variation between the studies appeared to be due to 2 trials of the ""atypical"" antidepressant mianserin, both in cancer patients, where, unlike in the rest of the trials, there was less significantly dropout on drug than on placebo (OR 0.26, 95% CI 0.13-0.54).
When we analysed the remaining 15 trials (trcicylics and SSRIs) without the 2 mianserin trials, there was no longer statistically significant heterogeneity. Grouped like this, antidepressants appear to produce a small but significant increase in dropout, OR 1.66, 95% CI 1.14-2.40.
Put another way, about 10 patients would need to be treated with antidepressants to produce one dropout from treatment which would not have occurred had they been given placebo (NNH 9.8, 95% CI 5.4-42.9).
Leaving the study prior to completion: comparing tricyclics and SSRI s.
Trials which randomised patients to either TCAs or SSRIs in the ohsycially ill depressed will be reviewed in subsequent editions of this review. However, it is of interest to compare the TCA-vs-placebo trials in this review with the SSRI-vs-placebo trials.
Nine studies used tricyclic antidepressants compared to placebo. A total of 62 of the 193 (32%) people randomised to a tricyclic failed to complete the study compared to 40 out of 184 (22%) randomised to placebo (OR 1.91 95% CI 1.19 to 3.07). Six studies compared SSRI s to placebo. 46 out of 139 people (33%) randomised to SSRI s failed to complete the study compared to 42 out of 140 (30%) randomised to placebo (OR 1.24 95% CI 0.71 to 2.15).
ANTIDEPRESSANTS COMPARED TO NO TREATMENT
The one trial which compared antidepressant with no treatment showed that patients on drug were significantly less likely to fail to improve (OR 0.29, 95% CI 0.09-0.91), but no more likely to drop out (OR 1.08, 95% CI 0.33-3.57). The same pattern of results is seen in the rest of the trials which compared antidepressants with placebo.
PHYSICAL ILLNESS CHANGES: disease-specific
Only 7 studies reported looking for changes in the physical disease. Antidepressants produced no change in immune function in HIV relative to placebo (2 studies) or no treatment (1 study). Relative to placebo, antidepressants produced no change in cardiovascular function in heart disease, in respiratory function in lung disease, or in vital signs or laboratory tests in cancer (1 study each). Nortriptyline produced worse control in diabetes.
These changes are detailed in the table. There appear to be no consistent trends.
PHYSICAL ILLNESS CHANGES: generic
Only 2 studies reported results from using generic quality of life scales, preventing any statistical summation across trials.
Markovitz (HIV) showed that antidepressants produced significntly higher group mean scores on the Karnofsky scale in the drug group than in the no treatment group, and Borson (lung disease) reported non-significant improvements on the Sickness Impact Profile on drug compared to placebo.i
ANTIDEPRESSANT EFFICACYThe results of the review provide evidence that antidepressants are more likely than placebo to cause improvement in depression symptoms in the short term in the depressed physically ill.
The grouping of all antidepressants together in a meta-analysis is supported by the lack of statistically significant heterogeneity for this outcome, and is in accord with previous reviews inidicating that different classes of drug have very similar efficacy against depression symptoms (e.g. 1).
The efficacy seen when the trials results are combined statistically is also consistent with the pattern of results seen if the results of individual trials are examined separately. In the latter case, 12/13 studies show ADs associated with greater chance of improvement than placebo; 5/12 studies are significant at the p <0.05 level, the remaining 7 are positive, but do not reach this conventional level of statistical significance.
ANTIDEPRESSANT EFFICACY BY DRUG CLASS
Regarding the two main types of antidepressant, the results are broadly consistent with previous reviews in the non-physically ill (e.g. ref 1), that TCAs and SSRIs have very similar efficacy. Although this was not a primary objective of this review, there was a non-significant trend in our review towards higher effectiveness in studies using TCAs than in those using SSRIs. Howewer, it would not be safe to conclude that the TCAs are more effective than the SSRIs in the depressed physically ill, as the differences are not statistically significant, and are based on comparisons between trials, and these comparisons are not themselves randomised. There might be other differences between the trials (e.g. in patients illnesses, the severity of their depression, etc) which are reponsible.
DROPOUT
Regarding dropout, there is no difference between drug and placebo if all antidepressants are grouped together.However, there was significant statistical heterogeneity between the studies so the above figure must be interpreted cautiously.
The 2 studies which used mianserin appear to be the ""outliers"" here, in that these were the only trials in which those given drug were significantly less likely to drop out than placebo. (This raises a possibility that there may be a specific effect of mianserin in depression in cancer, though this was not an a prioiri hypothesis; the possibility of blood dyscrasias with this drug and consequent necessity for blood tests make it unlikely to be drug of first choice ).
Mianserin is chemically distinct from the other antidepressants, being neither a tricyclic nor a selective serotonin reuptake inhibitor. Hence, it seems reasonable to examine the effect of removing these 2 trials from the analysis. If these 2 studies are removed from the comparison, there is no longer significant statistical heterogeneity between the remainaing studies (TCAs and SSRIs versus placebo). A small but significant increase in dropout is now apparent, OR 1.66 (95% CI 1.14-2.40).
DROPOUT BY DRUG CLASS
Regarding the two main types of antidepressant, the results are broadly consistent with previous reviews in the non-physically ill (e.g. ref 1) that TCAs and SSRIs have very dropout similar rates, though possibly slightly less on the latter. There was a greater tendency to dropout in the trials using tricyclic antidepressants compared to placebo (OR 1.91 95% CI 1.19 to 3.07) than in those comparing SSRIs to placebo (OR 1.24 95% CI 0.71 to 2.15).
Again, however, it would not be safe to conclude that the TCAs are more likely to cause dropout than the SSRIs in the depressed physically ill, as the differences are based on comparisons between trials, and these comparisons are not themselves randomised. There might be other differences between the trials (e.g. in patients illnesses, the severity of their depression, etc) which are reponsible.
It should also be noted that the same proportion of patients taking SSRI s and tricyclics failed to complete treatment (about a third). The apparent difference on this outcome measure is due to a different placebo response: 30% of those on placebo dropped out in the SSRI trials, but only 22% in the TCA trials.
PHYSICAL ILLNESS CHANGES: generic
Depression treatment would be hoped to lead to improvements in quality of life and function. It is therefore regrettable that only 2 studies reported results from using generic quality of life scales, preventing any statistical summation across trials.
Markovitz (HIV) showed that antidepressants produced significntly higher group mean scores on the Karnofsky scale in the drug group than in the no treatment group, and Borson reported non-significant improvements on the Sickness Impact Profile on drug compared to placebo.
PHYSICAL ILLNESS CHANGES: disease-specific
Depression treatment would in general not be expected to make major changes in established physical disease. The most striking finding was that nortriptyline worsened glycaemic control; SSRIs such as fluoxetine, which have a anorectic and hypoglycaemic effects, would be a better choice in the clinical situation. Antidepressants did not affect immune function in HIV/AIDS.
POTENTIAL PROBLEMS WITH THE EVIDENCE
Potential problems with the review process and with the trials themselves will be discussed separately below.
POTENTIAL PROBLEMS WITH THE REVIEW PROCESS: POSSIBILITY OF UNIDENTIFIED NEGATIVE TRIALS
First, all the included trials except one are small and positive. It is possible that other trials exist which are negative: such trials are known to be harder to identify, because they are more frequently unpublished, or are published in less accessible journals. It is possible that further trials will come to light after the publication of this review, which may be negative. However, given the strength of the effect found, it seems more likely that the conclusion of the review will be changed quantitatively than qualitatively.
Further releases of the Cochrane software (REVMAN) may permit the assessment of the likelihood existence of unidentified negative trials through the use of a funnel plot.
POTENTIAL PROBLEMS WITH THE REVIEW PROCESS: COMBINATION OF DIFFERENT DRUGS, DIAGNOSTIC METHODS AND DISEASES
Second, there is the question of our combining of statistical results across drugs, across methods of depression diagnosis and across different physical diseases. Although meta-analysis is frequently criticised on such grounds, we recognise that this does present a particular potential threat to the validity of the review in our case, and discuss it in detail here.
As has been noted above, the lack of statistical heterogeneity for inprovement and (except in 2 trials of mianserin in cancer) for dropout is consistent with the validity of the combining of the trials statistically regarding all these three parameters.
Combination across antidepressant classes (e.g. 2), and indeed across drug classes (2), has been done by other reviewers. We feel that combination of drug classes present the least potential threat to the validity of the review.
POTENTIAL PROBLEMS WITH THE REVIEW PROCESS: COMBINATION OF DIFFERENT DIAGNOSTIC METHODS FOR DEPRESSION
Regarding combination of results across trials using different diagnostic methods for depression, again the results across the trials are broadly similar whichever method was used, with no significant statistical heterogeneity. The effect seen is similar in the trials overall (0.37, 95%CI 0.27-0.51), or if the subset of 6 trials in which the diagnostic process was most rigourous is examined separetly (OR 0.35, 95% CI 0.22-0.55).
These 6 trials were of patients diagnosed as having DSM III or DSM IV major depressive disorder by structured clinical interview. Our review is therefore consistent with the idea that antidepressants may have a similar effect in physically ill patients diagnosed as depressed according to stringent reserach criteria as in such patients diagnosed by other means (including questionnaires and clinical methods).
It has been suggested that different diagnostic criteria should be used for the diagnosis of depressive illness in the physically ill. For example, that symptoms such as lack of energy, which are often associated with phsycial illness, should be dropped from depression criteria. Unfortunately, data from identified trials is not capable of answering the question of whether such variations in depression definition are associated with differential response to treatment.
POTENTIAL PROBLEMS WITH THE REVIEW PROCESS: COMBINATION OF DIFFERENT PHYSICAL DISEASES
Regarding combining trials in depression in different physical illnesses, we do not need to assume that depression is absolutely identical in all phsyical diseases for this combination to be valid. This would be intrinsically unlikely.
We only need to assume that it is reasonably similar. This assumption is consistent with international classifications such as DSM and ICD, which do not have separate categories for depression (or indeed for any mental disorder) in each of the myriad of different physical illness.
The lack of differnce in treatment response between stringently and less stringently diagnosed depressive illness within the review also counts against small potential differences in types of depression amongst different illnesses having a large effect on the outcome.
The main exception here would seem to be the cancer/mianserin instance, discussed above.
POTENTIAL PROBLEMS WITH THE TRIALS THEMSELVES
Most of the trials did not report how they randomised patients, a potential bias which may be associated with exaggerating the effects of treatment.
Also, most of the trials used observers to rate the outcome, rather than the patients themselves. The single trial which used a patient rated outcome showed a non-significantly positive treatment effect only. This may indicate an observation bias in favour of the drug, as trained observers are known to be able to discern antidepressants from placebo in a blinded trial more frequently than would be expected by chance.
Further, most of the studies did not examine the effect on the disease in question or the effect on overall function and quality of life.
REFERENCE
1) Song F et al. Selective serotonin reuptake inhibitors: meta-analysis of efficacy and acceptability. BMJ 306, 683-687, 1993.
2) Lima MS,Moncrieff J A comparison of drugs versus placebo for the treatment of dysthymia: a systematic review. Cochrane Library, 1998, Issue 3.
About 4 patients would need to be treated with antidepressants to produce one recovery from depression which would not have occurred had they been given placebo (NNT 4.2, 95% CI 3.2-6.4).
Antidepressants seem reasonably acceptable to patients, in that about 10 patients would need to be treated with antidepressants to produce one dropout from treatment which would not have occurred had they been given placebo (NNH 9.8, 95% CI 5.4-42.9).
The extrapolation of these results to clinical practice presents certain problems, however.
Not all low mood in the phsyically ill constitutes depressive illness. Episodes of transient emotional distress, including depressed mood, such as adjustment disorders, are common in the physically ill. However, persistent low mood can indicate depressive illness.
The existence, demonstrated in this review, of a cheap, readily available and effective treatment for depression should encourage detailed assessment of persistent low mood in the physically ill.
Regarding such assessment, the evidence does not suggest that standard criteria are inappropriate for the diagnosis of depression in the physically ill, but clearly they have to be interpreted in the context of the individual patient.
In view of the difficulty in diagnosis and unpredictability of response in the clinical setting, a trial of therapy may be indicated in doubtful cases.
Use of antidepressants should be considered in the physical illness who are diagnosed with depressive illness. The review indicates approximately equal effectiveness and acceptability, as judged by treatment completion, of TCAs and SSRIs.
More detail needs to be given on how samples are selected, and how they are defined as depressed.
Depression diagnosis by structured interview is regarded as a gold standard. The precise features best defining the syndrome could be investigated by examing the relation between treatment response and dropping symptoms such as fatigue, which can also be due to physical disease.
Clear descriptions of sample selection filters between study sample, study population and eventual target population for the drug should be given.
Although quick treatment response is of interest to patients and their carers, most existing trials are too short to judge whether long term effects are seen.
Regarding outcomes, function and quality of life should be measured as well as depression symptoms, using a generic measure (e.g.SF36, Dartmouth Co-op). Depression should be measured by at least one patient-rated and at least one observer rated scale.
Specific questions arise from the review .
Mianserin in cancer: does it have a specific action, and if so, do the advantages outweigh the necessity for blood tests to monitor the risk of blood dyscrasias?
Diabetes: could an SSRI be an effective treatment for depression without worsening diabetic control?
Medical Research Council, UK.
Dr Lustman generously supplied a preprint of his paper and Prof. Wilson additional information.
We are grateful to the Cochrane Library and Journal of Psychosomatic Research for allowing the parallel submission of this review.
References to studies included in this review
Cancer: Costa 85 (published data only) Cancer: van Heer. 96 (published data only) Diabetes: Lustman 96 (published data only) Head injury: Wroblew (published data only) Heart disease: Veith (published data only) Veith RC, Raskind MA, Caldwell JH, Barnes RF, Gumbrecht G, Ritchie JL. Cardiovascular effects of tricyclic antidepressants in depressed patients with chronic heart disease. N Engl J Med 1982; 306: 954-9. HIV/AIDS: Elliott (published data only) Elliott AJ et al. Randomized, placebo-controlled trial of paroxetine versus imipramine in depressed HIV positive outpatients. Am. J. Psychiatry, 155, 367; : -372, 1998. HIV: Markowitz (published data only) Markowitz JC, Klerman GL, Clougherty KF et al. Individual psychotherapies for depressed HIV-positive patients. Am J Psychiatry 1995; 152: 1504-9. Mrakowitz JC et al. Treatment of depressive symptoms in. HIV; : -positive p. HIV: Rabkin 94 (published data only) Rabkin JG, Rabkin R, Harrison W, Wagner G. Effect of imipramine on mood and enumerative measures of immune status in depressed patients with HIV illness. Am-J-Psychiatry 1994; 151(4): 516-23. HIV: Targ 94 (published data only) HIV: Zisook (published data only) Lung disease: Borson (published data only) Borson S, McDonald G, Gayle T, Deffebach M, Lakshminarayan S, VanTuinen C. Improvement in mood, physical symptoms, and function with nortriptyline for depression in patients with chronic obstructive pulmonary disease. Psychosomatics. 1992 Spri; 33(2): 190-201 1992. MS: Schiffer 90 (published data only) Schiffer R, Wineman N. Antidepressant pharmacotherapy of depression associated with multiple sclerosis. Am J Psychiatry. 1990 Nov; 147(11): 1493-7 1990. Phys. ill old Evans (published and unpublished data) Evans M et al. Placebo-controlled treatment trial of depression in elderly physically ill patients. Int. J. Geriatr. Psychiat. 12, 817; : -24, 1997. Phys. ill: Rifkin 85 (published data only) Rifkin A, Reardon G, Siris S et al. Trimipramine in physical illness with depression. J Clin Psychiatry. 1985 Feb; 46(2 Pt 2): 4-8 1985. Renal: Blumenfield (published data only) Stroke: AndersenG 94 (published data only) Andersen G, Vestergaard K, Lauritzen L. Effective treatment of poststroke depression with the selective serotonin reuptake inhibitor citalopram. Stroke. 1994 Jun; 25(6): 1099-104 1994;. Stroke: Lipsey 84 (published data only) Stroke: Reding (published data only) 1. Reding MJ, Orto LA, Winter SW, Fortuna IM, Di Ponte P, McDowell FH. Antidepressant therapy after stroke. A double-blind trial. Arch Neurol 1986; 43: 763-5. * indicates the major publication for the study References to studies excluded from this review Alcoff Alcoff J, Jones E, Rust P, Newman R. Controlled trial of imipramine for chronic low back pain. J Fam Pract. 1982 May; 14(5): 841-6. Alpsoy Alpsoy E, Ozcan E, Cetin L et al. Is the efficacy of topical corticosteroid therapy for psoriasis vulgaris enhanced by concurrent moclobemide therapy? A double-blind, placebo- controlled study. J-Am-Acad-Dermatol 1998; 38(2 Pt 1): 197-200. Andersen J Andersen J, Aabro E, Gulmann N, Hjelmsted A, Pedersen H. Anti depressive treatment in Parkinson s disease. A controlled trial of the effect of nortriptyline in patients with Parkinson s disease treated with L DOPA. Acta Neurol Scand. 1980 Oct; 62(4): 210-9 1980;. AndersenO Andersen OK, Bergsaker Aspoy J, Halvorsen L, Giercksky KE. Doxepin in the treatment of duodenal ulcer. A double-blind clinical study comparing doxepin and placebo. Scand-J-Gastroenterol 1984; 19(7): 923-5. Bacal Bacal B et al. Acao da doxepine em pacientes neuroticos com sintomalogia cardiaca. Rio Hospital 76, 225; : -234, 1969. Brown Bucknall Carrieri Carter Cater FS. Treatment of depression associated with chronic gastrointestinal and cardiovascular disorders. Psychosomatics, 9, 314; : -318, 1968. Caruso Chaplan Cunningham Cunningham LA. Depression in the medically ill: choosing .an antidepressant. J Clin Psychiatry 1994; 55 S: 90-7. Dam Dam M et al. Effects of fluoxetine and maprotiline on functional recovery in post-stroke hemiplegic patients undergoing rehabilitation therapy. Stroke, 27, 1211; : -1214, 1996. de Simone Driver Driver PS. Depression treatment in Parkinson s disease. DICP 1991; 25(2): 137-8. Evans DL Evans DL. Treatment of Depression in cancer patients is.associated with better life adaptation. Psychosomatic Medicine 50 72; : -76 1988. Evans W Evans W et al. The effects of antidepressant drugs on pain relief and mood in the chronically ill. Psychosomatics, 14, 214; : -219, 1973. Fedoroff Fedoroff JP, Robinson RG. Tricyclic antidepressants in the treatment of poststroke depression. J-Clin-Psychiatry 1989; 50 Suppl: 18-23; dis-6. Feinmann Feinmann C, Cawley R. Psychogenic facial. pain; : presentati-438, 1984. Fernandez Fernandez F et al. Effects of methylphenidate in HIV-related. depr; : a comparat-67, 1995. Ferrando Ferrando SJ, Goldman JD, Charness WE. Selective serotonin reuptake inhibitor treatment of depression in symptomatic HIV infection and AIDS. Improvements in affective and somatic symptoms. Gen-Hosp-Psychiatry 1997; 19(2): 89-97. Fischer Fischer P, Baas H. Therapeutic efficacy of R-deprenyl as adjuvant therapy in advanced Parkinsonism. J Neural Transm (Suppl) 25, 137; : -147, 1987. Frank Light R, Merrill E, Despars J, Gordon G, Mutalipassi L. Doxepin treatment of depressed patients with chronic obstructive pulmonary disease. Arch Intern Med. 1986 Jul; 146(7): 1377-80. Glick Gomez Perez Gomez Perez J, Nortriptyline and fluphenazine in the treatment of diabetic neuropathy. Pain,23,. 395; : -400, 1985. Goodnick Gottfries Gottfries CG et al. Effect of Citalopram, a Selective. 5; : -HT reuptak. Grassi Grassi B, Gambini O, Scarone S. Notes on the use of fluvoxamine as treatment of depression in HIV-1-infected subjects. Pharmacopsychiatry 1995; 28(3): 93-4. Grayburn Grayburn Davis W. Treating psychical disturbances induced by physical disorders. Psychosomatics, 9, 44; : -46, 1968. Grillage Gustafson Gustafson Y, Nilsson I, Mattsson M, Astrom M, Bucht G. Epidemiology and treatment of post-stroke depression. Drugs-Aging 1995; 7(4): 298-309. Hackett Hackett GI. Chronic muscle contraction headache: the importance of depression and anxiety. J Roy Soc Med 1987; 80: 689-91. Hameroff Hameroff SR, Weiss JL, Lerman JC et al. Doxepin s effects on chronic pain and depression: a controlled study. J-Clin-Psychiatry 1984; 45(3 Pt 2): 47-53. Hauser Hauser RA, Zesiewicz TA. Sertraline for the treatment of depression in Parkinson s disease. Mov-Disord 1997; 12(5): 756-9. Himmelhoch HIV: Mauri Huyse Johansson Johansson F, Von KL. A double blind controlled study of seratonin uptake inhibitor (Zimelidine) versus placebo in chronic pain patients. Pain. 1979 Aug; 7(1): 69-78 1979. Jokinen Katon Katon W, Sullivan M, Russo J, Dobie R, Sakai C. Depressive symptoms and measures of disability: a prospective study. J Affect Disord. 1993 Apr; 27(4): 245-54. Kishore Kumar Knobel Knobel M. The use of trazodone in psychosomatic medicine. Psychother. Psychosomat. 24 141; : -145, 1974. Koenig Koenig H, Goli V, Shelp F et al. Antidepressant use in elderly medical inpatients: lessons from an attempted clinical trial. J Gen Intern Med. 1989 Nov-; 4(6): 498-505. Kvindesal Kvindesal B et al. Imipramine treatment of painful diabetic neuropathy. JAMA, 251, 1727; : -1730, 1984. La Raja Laitinen Laitinen L. Desipramine in treatment of Parkinson s. dise; : -. Lakshmanan Lakshmanan M, Mion L, Frengley J. Effective low dose tricyclic antidepressant treatment for depressed geriatric rehabilitation patients. A double blind study. J Am Geriatr Soc. 1986 Jun; 34(6): 421-6. Lancaster Lancaster SM, Prout B, Pinto T, Anderson J, Schiff A. Influence of drug treatment on the irritable bowel syndrome and its interaction with psychoneurotic morbidity. Acta Psychiatr Scand. 1982 Jul; 66(1): 33-41. Lascelles Lascelles R. Atypical facial pain and depression. Br J Psychiatry. 1966 Jul; 112(488): 651-9. Lauritzen Lauritzen L, Bendsen B, Vilmar T, Bendsen E, Lunde M, Bech P. Post stroke depression: combined treatment with imipramine or desipramine and mianserin. A controlled clinical study. Psychopharmacology Berl. 1994 Feb; 114(1): 119-22. Light Light R, Merrill E, Despars J, Gordon G, Mutalipassi L. Doxepin treatment of depressed patients with chronic obstructive pulmonary disease. Arch Intern Med. 1986 Jul; 146(7): 1377-80 1986. Loldrup Loldrup D, Langemark M, Hansen HJ, Olesen J, Bech P. Clomipramine and mianserin in chronic idiopathic pain syndrome. A placebo controlled study. Psychopharmacology-Berl 1989; 99(1): 1-7. Macfarlane Maguire Manna Manna V, Bolino F, Di CL. Chronic tension type headache, mood depression and serotonin: therapeutic effects of fluvoxamine and mianserine. Headache. 1994 Jan; 34(1): 44-9 1994. Mantovani Mantovani, G et al. Evaluation of Multidimensional Instruments of Health Related Quality of Life in Elderly Cancer Patients undergoing 3 different psychosocial treatment approaches - a randomised controlled trial. Supportive Care in Cancer 1996 Vol 4, Issue 2, P129; : -140. Mauri Mauri M et al. A double-blind study on fluvoxamine vs. placebo in depressed HIV positive pati; : short term-201,1994. Max 1 Max M, Schafer S, Culnane M, Smoller B, Dubner R, Gracely R. Amitriptyline, but not lorazepam, relieves postherpetic neuralgia. Neurology. 1988 Sep; 38(9): 1427-32 1988. Max 2 Max M, Culnane M, Schafer S et al. Amitriptyline relieves diabetic neuropathy pain in patients with normal or depressed mood. Neurology. 1987 Apr; 37(4): 589-96. Max 3 Max M, Kishore KR, Schafer S et al. Efficacy of desipramine in painful diabetic neuropathy: a placebo controlled trial. Pain. 1991 Apr; 45(1): P69; Discu-3 1991. Max 4 Max M, Lynch S, Muir J, Shoaf S, Smoller B, Dubner R. Effects of desipramine, amitriptyline, and fluoxetine on pain in diabetic neuropathy (see comments). N Engl J Med. 1992 May 7; 326(19): 1250-6. Mayberg Meyers Miller Miller MD, Curtiss EI, Marino L et al. Long-term ECG changes in depressed elderly patients treated with nortriptyline. A double-blind, randomized, placebo-controlled evaluation. Am-J-Geriatr-Psychiatry 1998; 6(1): 59-66. Nappi Nappi G, Sandrini G, Granella F et al. A new 5 HT2 antagonist (ritanserin) in the treatment of chronic headache with depression. A double blind study vs amitriptyline. Headache. 1990 Jun; 30(7): 439-44 1990. Nyth Nyth A, Gottfries C. The clinical efficacy of citalopram in treatment of emotional disturbances in dementia disorders. A Nordic multicentre study. Br J Psychiatry. 1990 Dec; 157: 894-901 1990. Nyth 2 Nyth A, Gottfries C, Lyby K et al. A controlled multicenter clinical study of citalopram and placebo in elderly depressed patients with and without concomitant dementia. Acta Psychiatr Scand. 1992 Aug; 86(2): 138-45 1992. O Kane O Kane M et al. Fluoxetine in the treatment of obesetype 2 diabetics. Diabet. Med. 1994 332; : -. Okasha Okasha A, Ghaleb H, Sadek A. A double blind trial for the clinical management of psychogenic headache. Br J Psychiatry. 1973 Feb; 122(567): 181-3 1973. Onghena Onghena P, Van HB. Antidepressant induced analgesia in chronic non malignant pain: a meta analysis of 39 placebo controlled studies (see comments). Pain. 1992 May; 49(2): 205-19 1992. Panerai Pheasant Pheasant H, Bursk A, Goldfarb J, Azen S, Weiss J, Borelli L. Amitriptyline and chronic low back pain. A randomized double blind crossover study. Spine. 1983 Jul-; 8(5): 552-7 1983; or. Poulsen Rabkin Rabkin 2 Raffaele Raffaele, R;Rampello, L;Vecchio, I;Tornali, C;Malaguarnera, M. Trazodone therapy of the post-stroke depression ( 1996 ) Archives of Gerontology and. Geri; : 23; 217-220. Raskind Razavi Reifler Reifler B, Teri L, Raskind M et al. Double blind trial of imipramine in Alzheimer s disease patients with and without depression (see comments). Am J Psychiatry. 1989 Jan; 146(1): 45-9 1989. Ritchie Ritchie J, Truelove S. Comparison of various treatments for irritable bowel syndrome. Br Med J. 1980 Nov 15; 281(6251): 1317-9 1980. Robertson Robinson Robinson R, Parikh R, Lipsey J, Starkstein S, Price T. Pathological laughing and crying following stroke: validation of a measurement scale and a double blind treatment study (see comments). Am J Psychiatry. 1993 Feb; 150(2): 286-93 1993. Robinson 2 Robinson RG. Depression and antidepressants in patients with brain injury. 18th Collegium Internationale Neuro-Psychopharmacologicum. Clin Neuropharm, 15, Suppl. 1, Pt. A, 638A; : -639A, 1992. Saper Sarzi Sarzi Puttini P, Cazzola M, Boccassini L et al. A comparison of dothiepin versus placebo in the treatment of pain in rheumatoid arthritis and the association of pain with depression. J-Int-Med-Res 1988; 16(5): 331-7. Schifano Schifano F, Garbin A, Renesto V et al. A double blind comparison of mianserin and maprotiline in depressed medically ill elderly people. Acta Psychiatr Scand. 1990 Mar; 81(3): 289-94 1990. Schiffer Schiffer RB, Herndon RM, Rudick RA. Treatment of pathologic laughing and weeping with amitriptyline. N Engl J Med 1985; 312: 1480-2. Schulberg Scott Scott TF, Nussbaum P, McConnell H, Brill P. Measurement of treatment response to sertraline in depressed multiple sclerosis patients using the Carroll scale. Neurol-Res 1995; 17(6): 421-2. Selvini Selvini A, Rossi C, Belli C, Corallo S, Lucchelli PE. Antidepressant treatment with maprotiline in the management of emotional disturbances in patients with acute myocardial infarction: a controlled study. J-Int-Med-Res 1976; 4: 42-9. Shima Shima S. The efficacy of antidepressants in post-stroke depression. Keio-J-Med 1997; 46(1): 25-6. Silbeberg Silberberg D, Armstrong R. Tranylcypromine in Multiple Sclerosis. Lancet, October 23, 1965, 852; : -3. Sindrup Sindrup S et al. The selective serotonin reuptake inhibitor citalopram relieves the symptoms of diabetic neuropathy. Clin. Pharm. Ther. 52, 547; : -52, 1992. Sindrup2 Sindrup S et al. Clomipramine vs. desiprmaine vs. placebo in the treatment of diabetic neuropathy. Br J Clin Pharmacol 30, 683; : -691, 1990. Sindrup3 Sindrup S et al. The selective serotonin reuptake inhibitor paroxetine is effective in the treatment of diabetic neuropathy symptoms. Pain 42 135; : -44 1990. Small Stamenkovic Stamenkovic M, Schindler S, Kasper S. (Therapy of post-stroke depression with fluoxetine. A pilot project). Nervenarzt 1996; 67(1): 62-7. Steur Steur ENHJ, Ballering LAP. Moclobemide and selegiline in the treatment of depression in Parkinson s disease. J. Neurol. Neurosurg. Psychiat. 63, 547; : -558, 1997. Strang Sullivan Sullivan M, Katon W, Russo J, Dobie R, Sakai C. A randomized trial of nortriptyline for severe chronic tinnitus. Effects on depression, disability, and tinnitus symptoms (see comments). Arch Intern Med. 1993 Oct 11; 153(19): 2251-9 1993. Tan Tan R, Barlow R, Abel C et al. The effect of low dose lofepramine in depressed elderly patients in general medical wards. Br J Clin Pharmacol. 1994 Apr; 37(4): 321-4 1994. Teri Teri L, Reifler B, Veith R et al. Imipramine in the treatment of depressed Alzheimer s patients: impact on cognition. J Gerontol. 1991 Nov; 46(6): P372-7 1991. Turkington Turkington R. Depression masquerading as diabetic neuropathy. JAMA. 1980 Mar 21; 243(11): 1147-50 1980. Van Van MM, Dierick M, De MF, Vereecken A. Treatment of depressive anxiety states associated with psychosomatic symptoms. A double blind multicentre clinical study: mianserin versus melitracen flupentixol. Acta Psychiatr Belg. 1983 Sep-; 83(5): 525-39 198 3. Veeninga Vercoulen Vercoulen JHMM et al. Randomised double-blind placebo-controlled trial of fluoxetine in chronic fatigue syndrome. Lancet 1996; 347: 858-61. Volicer Volicer L, Rheaume Y, Cyr D. Treatment of depression in advanced Alzheimer s disease using sertraline. J Geriatr Psychiatry Neurol. 1994 Oct-; 7(4): 227-9 1994.Costa D, Mogos I, Toma T. Efficacy and safety of mianserin in the treatment of depression of women with cancer. Acta-Psychiatr-Scand-Suppl. 320. 1985:85-92. .
van Heeringen K, Zivkov M. Pharmacological treatment of depression in cancer patients. A placebo-controlled study of mianserin. Br-J-Psychiatry. 169. 1996:440-3.
Lustman PJ, Griffith LS, Clouse RE et al. Effects of nortriptyline on depression and glycemic control in diabetes: results of a double-blind, placebo-controlled trial. Psychosom-Med. 59. 1997:241-50.Lustman PJ, Griffith LS, Freedland KE, Clouse RE. The course of major depression in diabetes. Gen-Hosp-Psychiatry. 19. 1997:138-43.
Markowitz JC, Klerman GL, Clougherty KF, Spielman LA, Fishman B, Perry SW. TREATMENT OF DEPRESSED HIV-POSITIVE PATIENTS. American-Psychiatric-Association,-150th-Annual-Meeting,-San-Diego 1997.
Targ EF, Karasic DH, Diefenbach PN, Anderson DA, Bystritsky A, Fawzy FI. Structured group therapy and fluoxetine to treat depression in HIV-positive persons. Psychosomatics. 35. 1994:132-7.
Blumenfield M, Levy NB, Spinowitz B et al. Fluoxetine in depressed patients on dialysis. INT-J-PSYCHIATRY-MED. 27. 1997:71-80.
Lipsey JR, Robinson RG, Pearlson GD, Rao K, Price TR. Nortriptyline treatment of post-stroke depression: a double-blind study. Lancet. 1. 1984:297-300.
Brown C, Schulberg HC, Madonia MJ. Clinical presentations of major depression by African Americans and whites in primary medical care practice. J-Affect-Disord. 41. 1996:181-91.
Bucknall C, Brooks D, Curry P, Bridges P, Bouras N, Ankier S. Mianserin and trazodone for cardiac patients with depression. Eur J Clin Pharmacol. 1988; 33(6): 565-9 1988; (all 16).
Carrieri PB, Indaco A, Gentile S, Troisi E, Campanella G. S-Adenosylmethionine treatment of depression in patients with Parkinsons s disease. A double-blind, crossover study versus placebo. CURR-THER-RES-CLIN-EXP. 48. 1990:154-60.
Caruso I, Fumagalli M, Boccassini L et al. Treatment of depression in rheumatoid arthritic patients. A comparison of S-adenosylmethionine (Samyr) and placebo in a double-blind study. CLIN-TRIALS-J. 24. 1987:305-10.
Chaplan A. Doxepin in anxious/depressed outpatients with gastrointestinal symptoms: double-blind comparison of doxepin and perphenazine/amitriptyline. pp. 32-42. In: Mendels J, ed. Sinequan (doxepin HC1). Amsterdam, Excerpta Medica, 1975.
De Simone C, Catania S, Trinchieri V, Tzantzoglou S, Calvani M, Bagiella E. Amelioration of the depression of HIV-infected subjects with L-acetyl carnitine therapy. J-DRUG-DEV. 1. 1988:163-6.
Frank R et al, Antidepressant analgesia in rheumatoid arthritis. J Rheumatol. 1988; 15: 1642-1638
Glick R, Harrison W, Endicott J, McGrath P, Quitkin F. Treatment of premenstrual dysphoric symptoms in depressed women. J Am Med Wom Assoc. 1991; 46(6): 182-5.
Goodnick PJ, Kumar A, Henry JH, Buki VM, Goldberg RB. Sertraline in coexisting major depression and diabetes mellitus. Psychopharmacol-Bull. 33. 1997:261-4.
Grillage M. Neurotic depression accompanied by somatic symptoms: a double blind comparison of flupenthixol and diazepam in general practice. Pharmatherapeutica. 1986; 4(9): 561-70 1986.
Himmelhoch J, Schechtman K, Auchenbach R. The role of trazodone in the treatment of depressed cardiac patients. Psychopathology. 1984; 17 Suppl 2: 51-63 1984.
Mauri MC, Ferrara A, Fabiano L, Ricci C, Invernizzi G. A double blind study on fluvoxamine vs. placebo in depressed HIV positive patients: Short-term and perspective results. INTEGR-PSYCHIATRY. 10. 1994:199-201.
Huyse FJ, Zwaan WA, Kupka R. The applicability of antidepressants in the depressed medically ill: an open clinical trial with fluoxetine. J-Psychosom-Res. 38. 1994:695-703.
Jokinen K, Koskinen T, Selonen R. Flupenthixol versus diazepam in the treatment of psychosomatic disorders: a double blind, multi centre trial in general practice. Pharmatherapeutica. 1984; 3(9): 573-81 1984.
Kishore-Kumar R et al.: Desipramine relieves post-herpetic neuralgia. Clin Pharm. Therap. 1990: 47; 305-12.
La Raja MC, Virno F, Mechella M et al. Depression secondary to tumors in patients who underwent surgery for mammary carcinoma: psycho-pharmaceutical and psychotherapeutic care. J-Exp-Clin-Cancer-Res. 16. 1997:209-16.
Macfarlane J, Jalali S, Grace E. Trimipramine in rheumatoid arthritis: a randomized double blind trial in relieving pain and joint tenderness. Curr Med Res Opin. 1986; 10(2): 89-93 1986.
Maguire P et al. Treatment of depression in cancer patients. Acta Psychiatr. Scand. 1985: 72 ( suppl. 320 ); 81-4.
Mayberg HS, Parikh RM, Morris PLP, Robinson RG. Spontaneous remission of post-stroke depression and temporal changes in cortical S-2-serotonin receptors. J-NEUROPSYCHIATRY-CLIN-NEUROSCI. 3. 1991:80-3.
Meyers C, Vranckx C, Elgen K. Psychosomatic disorders in general practice: comparisons of treatment with flupenthixol, diazepam and sulpiride. Pharmatherapeutica. 1985; 4(4): 244-50 1985.
Panerai AE, Monza G, Movilla P, Bianchi M, Francucci BM, Tiengo M. A randomized, within-patient, cross-over, placebo-controlled trial on the efficacy and tolerability of the tricyclic antidepressants chlorimipramine and nortriptyline in central pain. ACTA NEUROLOGICA SCANDINAVICA. 82. 1990:34-8.
Poulsen D, Hansen H, Langemark M, Olesen J, Bech P. Discomfort or disability in patients with chronic pain syndrome. Psychother Psychosom. 1987; 48(1-4): 60-2 1987.
Rabkin JG, Rabkin R, Wagner G. Effects of fluoxetine on mood and immune status in depressed patients with HIV illness. J-Clin-Psychiatry. 55. 1994:92-7.
Rabkin JG, Wagner G, Rabkin R. Treatment of depression in HIV+ men: Literature review and report of an ongoing study of testosterone replacement therapy. ANN-BEHAV-MED. 18. 1996:24-9.
Raskind M, Veith R, Barnes R, Gumbrecht G. Cardiovascular and antidepressant effects of imipramine in the treatment of secondary depression in patients with ischemic heart disease. Am-J-Psychiatry. 139. 1982:1114-7.
Razavi D. et al. The effect of fluoxetine on anxiety and depression in cancer patients. Acta.Psychiatr.Scan. 1996: 94: 205-210.
Robertson M, Trimble M. The treatment of depression in patients with epilepsy: a double-blind trial. J. Affect. Dis.: 1985: 9; 127-136
Saper JR, Silberstein SD, Lake AEI, Winters ME. Double-blind trial of fluoxetine: Chronic daily headache and migraine. Headache. 34. 1994:497-502.
1. Coulehan JL, Schulberg HC, Block MR, Madonia MJ, Rodriguez E. Treating depressed primary care patients improves their physical, mental, and social functioning.. Arch-Intern-Med. 157. 1997:1113-20.2. Schulberg HC, Madonia MJ, Block MR et al. Major depression in primary care practice. Clinical characteristics and treatment implications.. Psychosomatics. 36. 1995:129-37.Brown C, Schulberg HC, Madonia MJ. Clinical presentations of major depression by African Americans and whites in primary medical care practice.. J-Affect-Disord. 41. 1996:181-91.
Small GW, Birkett M, Meyers BS, Koran LM, Bystritsky A, Nemeroff CB. Impact of physical illness on quality of life and antidepressant response in geriatric major depression. J-Am-Geriatr-Soc. 44. 1996:1220-5.
Strang RR. Imipramine in the treatment of parkinsonism: a double-blind placebo study. BMJ: 1965: 2; 33-34
Veeninga A, Westenberg H, Weusten J. Fluvoxamine in the treatment of menstrually related mood disorders. Psychopharmacology Berl. 1990; 102(3): 414-6 1990.