QUEST - Medically Ill, Co-Morbid Psychiatric Illness

• Management to reduce time in hospital in people with severe mental illness

• Shared Decision-Making Preferences of People with Severe Mental Illness

• Defense Mechanisms in Hemodialysis Dependent Patients

• Stimulants in Depressed Medically Ill Patients

• Nortriptyline and Medically Ill Cancer Patients

• Propofol Versus Midazolam in the Sedation of Critically Ill Patients

• SSRIs and Hyponatremia

• Methylphenidate effective as monotherapy for depression in the elderly, medically ill

 

 

---

AUTHOR
David Karol

Management to reduce time in hospital in people with severe mental illness

Chairman’s Rounds, October 1, 2007 

Citation: Burns T, Catty J, Dash M, Roberts C, Lockwood A, Marshall M. Use of intensive case management to reduce time in hospital in people with severe mental illness: systematic review and meta-regression. BMJ 2007;335;336-. Originally published online 13 Jul 2007; doi:10.1136/bmj.39251.599259.55 

Clinical Question: What independent variables in randomized control trials involving intensive case management can help predict mean patient days per month spent in the hospital? 

Background Information: The use of intensive case management and assertive community treatment for the seriously mentally ill has come under scrutiny in recent years due to inconsistent effects on the use of hospital care. Hypotheses for such inconsistent effects have included differences in health care system funding among different countries, differences in fidelity to the model of assertive community treatment, and differences in baseline hospital use among patients under intensive case management. 

Question Type/Study Design Type: Systematic Review and Meta-Regression 

Validity Criteria:

1)       Did the review explicitly address a sensible clinical question? Yes

2)       Was the search for relevant studies detailed and exhaustive? Yes, the authors searched the Cochrane central register of controlled trials, CINAHL, Embase, Medline, and PsychINFO databases from inception to January 2007.

3)       Were relevant studies likely to be omitted? No, they excluded trials that had inappropriate interventions and extremes of age. Disagreements between two authors were resolved by a third reviewer.

4)       Were the primary studies of high methodological quality? Unable to fully assess, though included studies were rated A or B via the Cochrane Collaboration Handbook

5)      Were the author’s assessments of the primary studies reproducible? Yes

6)      How homogeneous were the various study results? Not homogenous; the purpose of the study was to try to explain heterogeneity of results 

Methods:         
Source: Cochrane central register of controlled trials, CINAHL, Embase, Medline, and PsychINFO databases from inception to January 2007.

Inclusion Criteria: RCTs that compared intensive case management (caseload up to and including 20) with standard care (community mental health team or outpatient clinic) or low intensity case management (caseload greater than 20) in people with severe mental disorder living in the community. Severe mental disorder defined as schizophrenia or schizophrenia-like disorder, bipolar disorder, or depression with psychotic features. Furthermore, trials had to have data on the dependent variable on an intention to treat basis for more than half of the trial participants. 

Exclusion Criteria:
1) Experimental intervention was acute crisis team or if control condition was hospital admission, remaining in hospital, or alternate form of case management

2) Trials in which most participants were either under 18 or over 65 or had a primary diagnosis of organic brain disorder or learning disability 

Studies Screened vs. Accepted: See Figure 1. 29 eligible trials identified, of which eight were multicenter trials. Data were obtained from four out of eight of these multicenter trials, and disaggregated into individual centers, yielding a total of 52 centers for analysis. 

Statistics: Meta-regression analysis was performed using data from 29 trials (52 centers) in which predictions of the effect of each independent variable on the dependent variable were concluded. A meta-analysis was repeated for 42 centers that had baseline data on hospital use (in order to reduce type 2 error, this included only covariates that were found to have a significant association with the dependent variable in the first analysis). Sensitivity analyses were also performed. 

Outcomes Measured: Mean number of hospital days per month, measured on the basis of a 24 month follow-up period. In trials that did not report 24 month data, the dependent variable was calculated from the nearest available follow-up point. 

Main Results:

1)       As a center’s organization fidelity score increased, the mean number of hospital days per month decreased on the order of 0.44 fewer days per month per one point increase.

2)       No other covariates were significant, including team membership subscale scores.

3)       When baseline hospital use was analyzed (42 out of 52 centers), a direct correlation was seen between the number of baseline hospital days and a  reduction in mean future hospital days per month. This was a significant effect. 

Conclusions: While this was a technically difficult study to perform (as well as to read), the authors made a fairly good effort at acquiring the necessary data, including contacting the authors of the individual studies to collect missing data when appropriate. The main conclusions of the study were that team structure and organization, but not necessarily team membership and staffing, lead to reduction in hospital use among the seriously mentally ill. Furthermore, it concluded that assertive community treatment appears to work best with patients who have a high rate of hospitalization. This study suggests that ACT teams might best use their efforts to focus on structure and organizational issues within the team as well as to target populations of mentally ill who are frequent users of hospital resources. 

Synopsis: This systematic review of randomized controlled trials involving assertive community treatment looked at different variables within the trials that may affect frequency of patient hospitalization. Key findings included that fidelity to team structure and organization as defined by the IFACT subscale, as well as increased baseline patient hospitalization, are both associated with significant reductions in mean hospital days per month. 

Index of Fidelity to Assertive Community Treatment (IFACT) 

Team Membership:

1)       ratio of patients to staff

2)       total size of the team

3)       extent of psychiatric input

4)       extent of nursing input 

Team structure and organization, whether the team:

1)       is the primary source of care for its patients

2)       is situated away from the hospital

3)       meets daily

4)       shares responsibility for caseloads

5)       is available 24 hours/day

6)       has a team leader who is also a case manager

7)       offers unlimited time for its services

Return to top of the page

---

AUTHOR
Matt Connor

Shared Decision-Making Preferences of People with Severe Mental Illness 

Chairman’s Rounds, October 22, 2007 

Citation:  Adams J, Drake R, Wolford G.  Shared Decision-Making Preferences of People with Severe Mental Illness.  Psychiatric Services 2007; 58 (9): 1219-1221. 

Clinical question: What preferences do people with mental illness have for shared decision making, and how do they perceive their current decision-making distribution? 

Background info: Western biomedical ethics place autonomy as a fundamental principle, and a patient is the person best suited to evaluate the impact of a given treatment and/or side effect on his or her own life.  Involving the patient more fully in the decision-making process has been shown to lead to better outcomes, partially because it enhances adherence.  This has not been a well-researched area in mental health. 

Study type: Cross-sectional pilot study 

Methods: Questionnaires given to 30 people with “severe mental illnesses”  

  Setting: Community mental health center in New Hampshire over the months of July to October in 2004 and 2005

  Patients: The first 15 questionnaires were a convenience sample, and the next 15 were randomly selected from any of the center’s clients and pooled with the first.  The groups were essentially equal except that the second 15 were more likely to have education past the high school level and less likely to have a diagnosis of schizophrenia, but a table is not reproduced, so this is an area where the reader has to trust the authors.  In the pooled group, mean age was 47 years.  14/30 were male, 10/30 reported having a schizophrenia spectrum diagnosis, 12/30 were unemployed, 22/30 had completed high school, and 18 of those 22 had education beyond high school.

  Tools:  Autonomy Preference Index Decision-Making scale (API-D) – 6 questions about autonomy in decision making, scaled 0 (completely passive) to 100 (utter autonomy)

              Control Preferences Scale (CPS) – A scale of 1 (“mostly my decision”) to 5 (“mostly my doctor’s decision”) applied to a few decision areas as seen Table 1

              Open-ended questions to confirm validity of the questionnaire responses (no report on who asked the questions or whether these were standardized) 

Outcome measures: CPS rating and association with age, sex, education, substance use (self-report), and diagnosis of schizophrenia (self-report); API-D results 

Results:  API-D result was 51 (SD 9), meaning equal collaboration, with no difference in relation to age, sex, education, substance use, or diagnosis of schizophrenia based on t-test.

   CPS results are in Table 1, and it is notable that 1 of the 30 patients did not fill out the half of the survey regarding his or her perceived role.  Scores of 1 or 2 were considered “autonomous,” 3 was “collaborative,” and 4 or 5 was “passive.”  The only place where perceived roles were significantly different (based on chi square) from preferred roles were concerning new medications and choice of practitioner.  There was also report that clients were less likely (by Wilcoxon ranked-sum test) to prefer a passive role in psychiatry as compared to in a general medical setting, but this was not presented in the table. 

Conclusions:  The API-D scores in this pilot study are consistent with other studies in European populations of the mentally ill.  Compared to patients with diabetes, these populations score higher on the API-D, meaning they prefer more autonomy.  In the CPS portion of the surveys, the authors note that though there are differences within the population, their mental health clients appeared to prefer active and collaborative roles, especially when it comes to psychotropic medications. They also subtracted the percent of clients who report preferring a passive role (23%) from the clients who report perceiving a passive role (62%) (This is mis-stated in the text) to conclude that 39% of clients would like a more active role than they currently experience.  This is consistent with results from another qualitative study, the commission on mental health, and the board on health care services. 

Weaknesses:  The paper itself points out that the sample size was very small, so the statistics lacked power.  The authors also acknowledge the relatively high education and employment levels in their population and point out that the study did not address capacity.

   I noted several other major weaknesses in this paper.  Due to format constraints, the authors could not publish a table to prove the equivalence between the first 15 subjects and the second 15, so the readers have to take their word on the few disclosed differences and assume that they compared every other relevant point. Though the authors compare the two halves of their sample pool as being roughly equal at baseline, they do not verify this equality by comparing end-results.  The patients all have “severe mental illness” but a list of diagnoses is not provided.  They also allow the patients to self-disclose for demographics of substance use and mental-health diagnosis when the study could conceivably have included chart review for Axis 1 diagnosis and a urine screen for active substance use.

   Also, the table displayed allows CPS scores of 1 and 2 to be pooled and scores of 4 and 5 to be pooled when it might have been more appropriate to display five columns or to use a scale with six possible answers.

   My main concern, however, was that the data was not connected.  In the CPS portions of the surveys, patients reported perceived and preferred roles, but there was no measure of difference between these two.  They assessed difference by subtracting the number of people who desired passivity from the number perceiving passivity, but we have no proof that this is an inclusive group.  What if a patient feels like he makes all the decisions in his care but would prefer to make none of them?  Their conclusions would have been more applicable if the data analysis had included something as simple as a data point reflecting difference between perceived and desired roles.

Summary: This paper is deeply flawed, but it brings up a good topic – the level of autonomy perceived by the patient is not always the level desired, and this may have an impact on adherence.  In an inpatient setting, patients are probably more likely to be incapable of decision-making, but in an outpatient world, it does behoove the clinician to figure out what role the patient wants to play in the relationship and to honor that as practicality allows.  Patients should always receive our respect.

Return to top of the page

---

AUTHOR:
Dan McCabe 

Defense Mechanisms in Hemodialysis-dependent Patients

Chairman’s Rounds, November 13, 2006

Citation: Zoccali R et al. Defense mechanisms in hemodialysis dependent patients. Clinical Nephrology. 2006; 65, 119-123

Clinical Question: What tools do patients use to accept chronic treatments like hemodialysis, and what psychological factors inhibit them from accepting chronic treatments?

Clinical Case or Background Info: 71yo Female with Chronic renal insufficiency who has resisted going on Dialysis despite advice of her physicians for 5 years.  Now with complete kidney failure and forced to accept maintenance hemodialysis.

Study Design Type: Case Control
     Setting: Dialysis unit of the Policlinico Universitario of Messina.
     Population: 60 outpatients, selected consecutively, on chronic hemodialysis and without major psychiatric comorbidity. Sex, age, and duration of dialysis varied widely.  Patients were subdivided by length of dialytic treatment into 21 patients with less then 5 years and 19 patients with more then 10 years.  Control Sample: 50 age and sex matched healthy individuals without major psychiatric co morbidity.

     Exposure: Long-term dialysis

Assessment and Outcomes: All subjects were assessed using the Defense Mechanism Inventory (DMI) which measures the frequency of usage of five major groups of defense mechanisms.  The test consists of 10 brief stories, 2 for each “conflict area” (authority, independence, femininity/masculinity, competition, unexpected events), followed by 4 questions: What would your actual reaction be?  What would you impulsively (in fantasy) want to do?  What thought might occur to you?  How would you feel and why?  Five choices of answer are given, one for each of the defense styles being measured.  The DMI identifies 5 defensive styles: Turning against object (TAO: displacement and identification with the aggressor), Projection (PRO: attribution of negative characteristics or intent to an external object), Principalization (PRN: rationalization, intellectualization and isolation), Turning against self (TAS: masochism), Reversal (REV: denial and repression)

Main Results: See Table 1.  Mean scores on DMI were all in the normal range for both groups.  Comparing the two groups, HD patients had significantly higher scores on REV and lower scores on TAO.  There were no significant differences in DMI scores between the two subgroups of HD patients.
Conclusions: This was a relatively small study using DMI which previously has had limited use in patients affected with organic illness.  Patients on hemodialysis in this study were shown to use TAO less. This may be because displacement defense mechanisms such as aggressive expressions directed toward the environment might reduce the possibility of receiving emotional and empathic support from the social network and family members that these patients are so dependent on.  However patients on HD were found to have higher uses of REV defense mechanisms including denial.  Despite the negative connotations of denial other studies have demonstrated the usefulness of denial for patient’s coping with debilitating disease.  Short and Wilson demonstrated in 1969 that as dialysis patients transitioned from pre-dialysis to outpatient dialysis to medical complications associated with their kidney failure that their use of repression increased and their anxiety levels declined.  They further theorized that denial was useful and necessary for the dialysis patient but counter productive for the family and medical team. One might expect that the longer patients are on hemodialysis, the more insight into their defense mechanisms they might gain, but this study seems to indicate this may not occur.

Return to top of the page

---

Author:
Juandolyn Peters

Stimulants in Depressed Medically Ill Patients

Clinical Question:
Are stimulants an appropriate/effective pharmacotherapy for depressed medically ill patients? 38 year old female with metastic colon cancer and depression with poor oral intake.

Search Methods: (non stated)

Article/Results: 
Fernandez, F; Adams, F.; Holmes, V: Levy, J; Neidhart, M.  Methylphenidate for Depressive Disorders in Cancer Patients.  Psychosomatics.  1987. 28 (9): 455 - 461.

Method:
Trial of methylphenidate in 30 depressed cancer patients who were selected from a group of 100 consecutive outpatients referred to M.D. Anderson Hospital and Tumor Institute in Houston, TX.  The patients were selected from a group of 45 determined to meet criteria for affective d/o based on severity of symptoms (suicidal/anorexic with severe weight loss requiring immediate intervention), considered incapable of tolerating other antidepressants due to side effects (elderly or multi system diagnosis or” organic” mental d/o or 2nd - 3rd degree AVB/tachyarrythmia/cardiac output from adriamycin toxicity). 

7/30 had previous psychiatric history, but only one with depression (others: substance abuse {3}, factitious d/o {1}, personality d/o {1} and psychosexual dysfunction {1}).

Age: 30 - 99

Karnofsky performance status of all was <40% (i.e. capable of only limited self care, confined to bed/chair for at least 50% of waking hours)

Of remaining 70 consecutive pts, 16 had affective d/o (dysthymia or atypical depressive d/o by DMS-III).  These were started on conventional antidepressant therapy based on family history with positive response to selected antidepressant. 

Diagnoses: 
Major depression (4)
Adjustment d/o with depressed mood (14)
Dementia and depressed mood (3)
“Organic” affective syndrome (6 – 5/6 were on Vincristine)

Treatment:
Methylphenidate 10mg TID to start; up to 80mg/day by the end of 2 wks in 5 – 10mg increments after 2 – 3days’ observation.

Results:

• 77% marked or moderate improvement; 4/46 minimal improvement 3/46 – no improvement

• 19/30 treated mean of 38 days; 11/30 treated one year and sustained mod-marked improvement

• All responders had increased appetite. Response was rapid in general.

Critical Review: 

• Sample selection: Non-random, small sample size, not controlled/blinded

• Results: Authors rated patients for improvements themselves; CGIS is not specific to depression

Clinical Recommendations:  
Psychostimulant trial may be appropriate for controlling depressive symptoms, particularly in terminally ill but the evidence remains weak. Controlled RCT s is needed. RCT comparison to antidepressants is needed.

Reviewed by: 
Ranga Krishnan MB., ChB., Duke University Medical Center's Department of Psychiatry and Behavioral Science, 9/2000

Return to top of the page

---

Author:
Drew Barzman

Nortriptyline and medically ill cancer patients

Clinical Question:
What is the evidence for or against using Nortriptyline in patients with cancer or who are medically ill?

Search Method and:
1) There were no formal studies that have investigated this question in cancer patients.  There were a few studies that looked at the use of TCA's in relieving pain rather than depression in patients with cancer.

2). For the medically ill, there was one retrospective study on TCA’s, which was very limited by the numbers of subjects and its design

Articles/ Results:
Roose, S.P. et. al.  Comparative Efficacy of Selective Serotonin Reuptake Inhibitors and Tricyclics in the Treatment of Melancholia. American Journal of Psychiatry.  151; 12:1735-1739.

Methods:
Patients that were included had cardiac disease and major affective disorder.  They were inpatients in a depression research unit with Hamilton scores that were higher than 18.  These subjects also needed to meet the cardiac inclusion criteria (pg. 1736). Patients who met the inclusion criteria were admitted for a 2-week placebo period and cardiovascular testing.  If their Hamilton score was still higher than 18 after these two weeks, the patient entered the study. This study used a parallel group design with a non-equal random subject assignment of 3 to 1 in favor of fluoxetine.  The first group included patients who were treated with fluoxetine in the first drug trial.  Their dosing schedule was 20mg/day for the first 2 weeks and 40mg/day for week 3.  If necessary, the dose was increased to 60mg/day for 4 more weeks. Patients had completed the study after 6 weeks of treatment.  The second group included patients who had been treated with Nortriptyline in the last three studies and in the current study.  After therapeutic levels were achieved with a titration protocol, the patients had completed the study after 4 weeks.  Response was defined as returning to baseline functioning, ability to leave the hospital for 2 weeks without requiring dosage adjustments, and the Hamilton score was less than 8.  Average age was 73 for the fluoxetine group while it was 70 for the Nortriptyline group.  The mean pretreatment Hamilton score was 26 and 28 for the for the fluoxetine and Nortriptyline group, respectively.

Results:
34 of 42 patients were able to complete their trial of Nortriptyline.  Six were nonresponders and eight dropped out.  28(82%) of these 34 had responded.  For melancholic completers, 20 of the 24 (83%) had responded.  The intent to treat response rate was 67% (28 out of 42). 18 of the 22 patients completed their trial of fluoxetine.  13 were non responders while four dropped out.  Only five of the 18 had responded.  One of the 10 patients with the melancholia subtype had responded.  The intent to treat response rate was 23% (5 of 22).  18% dropped out compared to 19% of the Nortriptyline group.  A logistic regression confirmed a significant difference while controlling for melancholia, age, and pretreatment Hamilton scores.

Critical Review:
This is a poor study:
1) Nonrandom assignment.  There may have been differences among the groups that affected outcome.  

2) There was a retrospective analysis of previously collected data from past studies.

3) There were a limited number of subjects.

4) From table 2, it appears that the fluoxetine group had a  higher percentage of patients with Hamilton  Depression Scores of 30.  From Table 1, 40% had previous treatment for the current episode in the Nortriptyline group versus 32% in the fluoxetine group.

12 double blind, placebo-controlled trials have shown superior responses of tricyclics for the elderly.  Hence, the elderly may respond differently than a younger, medically ill population.

Clinical Recommendations: 
This is a poor study that showed a statistical significant.  Difference in efficacy between Nortriptyline and Fluoxetine in the treatment of hospitalized elderly patients with unipolar major affective disorder as well as melancholia and cardiovascular disease.  This study does not provide strong evidence for or against the use of the Nortriptyline in patients with cancer or the medically ill.  However, it suggests that Nortriptyline can be considered a potentially effective treatment for patients with depression and cardiovascular disease. A later study by Roose suggests that Paroxetine may be a better option than Nortriptyline.

Roose SP, Laghrissi-Thode F, Kennedy JS, Nelson JC, Bigger JT Jr, Pollock BG, Gaffney A, Narayan M, Finkel MS, McCafferty J, Gergel I.

Comparison of Paroxetine and Nortriptyline in depressed patients with ischemic heart disease. JAMA. 1998 Jan 28;279(4):287-91.

Reviewed by: 
Ranga Krishnan MB., ChB.,  Duke University Medical Center's Department of Psychiatry and Behavioral Science, 9/2000

Return to top of the page

---

Comparative study of propofol versus midazolam in the sedation of critically ill patients: Results of a prospective, randomized, multicenter trial

Chamorro, C., De Latorre, F. J., Montero, A., et al. Critical Care Medicine 24 (6): 932-939

Summary: Multicenter, prospective, randomized, open-labeled comparison of propofol and midazolam in sedation. Pt are from the 9 Spanish ICUs, mechanically ventilated, hemodynamically stable, and required sedation > 2days. P and M were infused and titrated to achieve "desired level of sedation". Level of sedation was obtained according to a sedation scale, q10’ x 6, then @6h, then q12h, and @discontinuation, up to 5days.

1. Are the results valid ?

1. Randomization ?: yes, by random numbers table, duration of study is 2-5 days
2. All pts were properly accounted for and analyzed in the group they were randomized to at the end of the study, with complete f/u (including death in 1/50 in P and 2/48 in M)

    3.    Blind ? No (you can’t due to the different look of the 2 meds)

4. groups were similar at the start of the study: age, gender, wt, h/o chronic dis (with detailed stratification), mortality prediction model score
5. aside from the experimental intervention, were the groups treated equally ?: Yes including the need for morphine and m relaxant

1. Results

1. Effectiveness (effective=4, acceptable=3, ineffective=<3)-- how large was the treatment effect? During the first hr: 66% in P (with 291 assessments) vs 69.1% in M (with 285 assessments); after the first hr: 76.5 % in P (with 332 assessments) vs 66.2% in M (with 355 assessments)

		1st hr			After 1st hr
	E=4	Accp=3	IE=<3	E=4	Accp=3	E=<3
Propofol	66%	27.8%	6.2%	76.5%	20.5%	3%
Midazolam	69.1%	22.5%	8.4%	66.2%	26.2%	7.6%

2. How precise was the estimate of the treatment effect

ARR= 10.3%, RRR= 13.5%, The 95% confidence interval for the absolute risk reduction (increase in response rate) is –7.5% to 28.1%

1. Will the results help me in caring for my patients ?

1. Can the results be applied to my pt care: Yes
2. Were all clinically important outcome considered ? Yes
3. Are the likely treatment benefits worth the potential harms and costs: yes, although the cost is not evaluated.

In the population of the critically ill pt, propofol is a effective and safe alternative for sedation, with some advantages, such as short duration of action and high effectiveness, over the conventional benzodiazepines.

Return to top of the page

---

SSRIs AND HYPONATREMIA

Bottom Line: SSRIs are more likely to cause hyponatremia than other antidepressants (ORadj 3.9 ) or other drugs (ORadj 3.5-4.0); patients over 65 may be at increased risk.  But the data are from retrospective studies.  The mechanism is SIADH.

Search: Medline: (Serotonin Uptake Inhibitors or Antidepressive Agents) and (Hyponatremia or Inappropriate ADH Syndrome)
          - Retrospective case-controls

References: (1) Movig KLL, et al.  Association between antidepressant drug use and hyponatraemia: a case-control study.  Br J Clin Pharmacol 2002;53:363-369.
- (2) Kirby D, et al.  Hyponatraemia in elderly psychiatric patients treated with Selective Serotonin Reuptake Inhibitors and venlafaxine: a retrospective controlled study in an inpatient unit.  Int J Geriatr Psychiatry 2002;17:231-237.
- (3) Movig KLL, et al.  Serotonergic antidepressants associated with an increased risk for hyponatremia in the elderly.  Eur J Clin Pharmacol  2002;58:143-148.  
- (4) Siegler EL, et al.  Risk factors for the development of hyponatremia in psychiatric inpatients.  Arch Intern Med 1995;155(9):953-957.  
- (5) Wilkinson TJ, et al.  Incidence and risk factors for hyponatremia following treatment with fluoxetine or paroxetine in elderly people.  Br J Clin Pharmacol 1999;47:211-217. 
- (6) Spigset O, Hedenmalm K.  Hyponatremia in relation to treatment with antidepressants: a survey of reports in the World Health Organization data base for spontaneous reporting of adverse drug reactions.  Pharmacotherapy 1997;17(2):348-352.  

Background: SIADH is 
          - Hyponatremia 
          - Low serum osmolarity
          - Inappropriately concentrated urine osmolarity

- Potential risk factors for hyponatremia include
          - Age > 65
          - Female 
          - Summer

Drugs	Diagnoses
Cardio meds	Renal disease
Diuretics	Hyperkalemia
Thiazides	DM
Loop diuretics	Cardio disease
ACE-I	HTN
Ca channel blockers	CHF
Nitrates	CAD
Beta blockers	MI
Psych meds	Angina
Antipsychotics	Pulm disease
Benzodiazepines	COPD
Antiepileptics	Lung CA
Analgesics	Emphysema
NSAIDs	Tobacco
Peptic ulcer drugs	Malignancy
PPI	Endo disease
H2 blockers	Hypoaldosteronism
	Hypothyroidism

Clinical Case: 59-yo male admitted for SOB.

Pulm: Restrictive/obstructive

lung disease: O2
                                        Prednisone
                                        Albuterol/atrovent
                                        Combivent
                                        Salmeterol
                                        Flunisolide
                                        Codeine
- Sleep apnea/obesity: CPAP

CV: CHF/CAD/HTN: ASA
                                      Lisinopril
                                      Furosemide
                                      Metolazone
                                      Spirinolactone
      - Afib/Aflutter: Heparin
                                  Coumadin
      - HL: Simvastatin
                Niacin

Renal: CRI: IVF
          - Hyperkalemia: K-dur

ID: SA bacteremia: Vancomycin

Endo: DM2: Glipizide
                      SSI

GI: Constipation/nausea: Docusate
                                             Lactulose
                                             Metoclopramide
   - GERD: Rabeprazole

GU: Impotence: Testosterone

M-S: Gout: Allopurinol

Derm: Rash: Fexofenadine

Psych: Depression: Prozac

                                 Sertraline

Date	Na	K	AD
01/03/02			Prozac 10
12/19/02	136		Prozac 10
01/28/03	136		Prozac 10
03/20/03	141		Prozac 10
04/04	132		Prozac 10
04/05	135		Prozac 10
04/06	134		Prozac 10
04/07	134		Prozac 10
04/08	132		Prozac 10
04/09	131		Prozac 10
04/10	133		Sert 50
04/11	132		Sert 50
04/12	132		---
04/13	131		Sert 100
04/14	128	5.5	Sert 100
Serum osm 296 Urine osm 360 (Na36 K38 Cl31)
04/15	132, 127	5.9, 6.5	---
04/16	136	5.8	---
04/21	132		---

(1)

Methods: Design: Retrospective case-control
- Setting: 2 teaching hospitals and large MHC
- Patients:

Na measured: 39 071
Na < 131:         1 391
Cases	Controls
Na<131 + AD: 29	Na>135 + AD: 78

- Inclusion: In- and outpts in database 6/97-6/99
- Exclusion: Younger than 18; 130<Na<136
- Matching: Same ward and admission week
- Outcomes: Primary: SSRI use
          - Secondary: drugs and diagnoses associated with hyponatremia
- Analysis: t-test, Chi-square, logistic regression, stratified analysis, interaction analysis

Validity: No dose-response or timing (drug d/c?)
- Patients are roughly same age as Case pt
- Observer bias: both paper and computer records for morbidity data collection
- Selection bias: controls from same ward + week
- Sample size: only 107

  Results: Primary outcome:

	Na < 131	Na > 135
SSRI	22/29 (76%)	38/78 (49%)
Non-SSRI AD	7/29 (24%)	40/78 (51%)

- OR 3.3 (CI 1.3-8.6)

- Secondary: adjusted OR 3.9 (CI 1.2-13.1)
          - K > 5: adj OR 24 (CI 2.0-283)
          - Age > 65: adj OR 6.3 (CI 1.0-41)
- Interaction: SSRI + diuretic: OR 8.4
          - SSRI + diuretic + >65: OR 13.5

(2)

Methods: Design: Retrospective case-control
- Setting: Acute psychogeriatric unit
- Patients:

Depressed pt and next pt: 218
Na measured:                    199
Cases	Controls
SSRI or venla: 74	Other meds: 125

- Inclusion: Inpatients on unit 1997-98; any Na
- Exclusion: Younger than 65; no Na level
- Matching: Non case depressed pt or next pt on ward
- Outcomes: Primary: Na < 135
          - Secondary: drugs and diagnoses (with severity) associated with hyponatremia
- Analysis: t-test, Chi-square, logistic regression, stratified analysis, interaction analysis

Validity: No dose-response or timing (drug d/c?)
- Patients are older than Case pt
- Observer bias: only paper records for morbidity data collection
- Selection bias: depressed pts but also next pt
- Sample size: 199

  Results: Primary outcome:

	SSRI or venla	Other meds
Na < 135	29/74 (39%)	13/125(10%)
Na > 134	45/74 (61%)	112/125(90%)

- OR 5.6 (CI 2.6-11.6)

- Secondary: adjusted OR 3.5 (CI 1.4-8.9)
          - Medical illness severity: adj OR 1.9
          - Depression diagnosis: adj OR 3.6
          - Thiazide: adj OR 2.8
- Insignificant: age, non-thiazide ‘causal’ drugs

(3)

Methods: Design: Retrospective case-control
- Setting: Hospitals in 8 Dutch cities
- Patients:

Database:                                     320 000
Hyponatremia or SIADH admission: 282
Cases	Controls
HypoNa inpt: 203	NormoNa outpt: 608

- Inclusion: In- and outpts in database 1/90-12/98
- Exclusion: Younger than 18; not case or match
- Matching: Birth year, sex, area; normoNa outpt
- Outcomes: Primary: SSRI, venla, clomipramine use
          - Secondary: drugs (including TCAs) and diagnoses associated with hyponatremia
- Analysis: Chi-square, logistic regression, stratified analysis, interaction analysis

Validity: Timing but no dose-response |
- Patients are older than Case pt
- Observer bias: only computer records for morbidity data collection
- Selection bias: HypoNa diagnosis; outpt normoNa
- Sample size: 811

Results: Primary outcome: 

	HypoNa adm	NormoNa out
SSRI/ven/clom	10/203 (5%)	8/608 (1%)
Other meds	193/203 (95%)	600/608 (99%)

                - OR 3.9 (CI 1.5-10.0)

- Secondary: adjusted OR 4.0 (CI 1.3-11.8)
          - Thiazide: adj OR 3.2
          - PPI: adj OR 2.3
          - CHF: adj OR 3.3
- Time to admit: median 10 days; 9/10 in 12 days

(4)

Methods: Design: Retrospective case-control
- Setting: Tertiary hospital’s inpt psych service
- Patients:

All psych inpatients over 3 years
Charts reviewed: > 90%
Cases	Controls
Na < 130: 64	Na > 134: 192

- Inclusion: Inpts with reviewed charts 1988-90
- Exclusion: 129<Na<135; not d/c in 1mo of hypoNa
- Matching: Randomly picked d/c in same month
- Outcomes: Primary: fluoxetine
          - Secondary: drugs (including TCAs) and diagnoses associated with hyponatremia
- Analysis: t-test, Chi-square, logistic regression, stratified analysis, interaction analysis

Validity: No dose-response or timing (drug d/c?)
- Patients match age of Case pt
- Observer bias: only paper records for morbidity data collection
- Selection bias: tertiary inpts; controls d/c same mo
- Sample size: 256

  Results: Primary outcome:

	Na < 130	Na > 134
Fluoxetine	12/64 (19%)	7/192 (4%)
Other meds	52/64 (81%)	185/192 (96%)

                - OR 6.1 (CI 2.3-16.3)

- Secondary: adjusted OR 21.4 (CI 5.3-86.9)
          - Hyperkalemia: adj OR 19.1
          - Diuretic: adj OR 8.2
          - TCA: adj OR 4.9
- Insignificant: age  

(5)

Methods: Design: Retrospective case-control
- Setting: Hospital’s inpt and outpt elder rehab
- Patients:

SSRI (fluoxetine or paroxetine): 845
Na < 130:                                      42
Cases	Controls
SSRI,Na<130 linked: 14	SSRI + Na>134: 56

- Inclusion: In- and outpts in database over 3.5y
- Exclusion: Age<65, CHF, 129<Na<135
- Matching: Randomly picked SSRI + Na>134
- Outcomes: Primary: incidence

- Secondary: drugs and demographics associated with hyponatremia
- Analysis: t-test, Chi-square, logistic regression, interaction analysis

Validity: Timing and attempted dose-response
- Patients are older than Case pt
- Observer bias: computer records for drugs and demographics
- Selection bias: “linked” hypoNa and SSRI use
- Sample size: mere 70

Results: Primary outcome:

          - 4.7 hypoNa per 1000 SSRI per year
          - 6.3 hypoNa per 1000 fluox per year
          - 3.5 hypoNa per 1000 parox per year

  - Secondary: Age: adj OR 1.1
          - Weight: adj OR 0.9
- Time to hypoNa: median 14 days; 79% in 21 days

(6)

Methods: Design: Retrospective case-control?
- Setting: The world
- Patients:

Humanity
Collaborating Centre database: many
Cases	Controls
HypoNa + AD: 668	Other SE + AD: 68 269

- Inclusion: Hyponatremia or SIADH plus AD reported in WHO database 1968 to 1993
- Exclusion: Incomplete information
- Matching: All other adverse effect reports
- Outcomes: Sex, age
          - Season
          - Time to hyponatremia, dose
- Analysis: t-test, Chi-square

Validity: Timing and dose-response
- Patients match age of Case pt
- Observer bias: disparate computer records for time to hyponatremia diagnosis
- Selection bias: Diagnosed, reported hypoNa + AD
- Sample size: just 68 937  

Results:

	HypoNa+AD	OtherSE+AD	P
Women	78%	69%	<.0005
Age	67y	49y	<.0001
Summer	31%	25%	<.02

- Time to hypoNa: 51% in 14 days; 75% in 30 days
- Insignificant: dose

Return to top of the page

Clinical question: Is methylphenidate effective as monotherapy for depression in the elderly, medically ill patient?

Citation:  Wallace Amy, Kofoed Lial, West Alan.  Double-blind, Placebo-Controlled Trial of Methylphenidate in Older, Depressed, Medically Ill Patients.  American Journal of Psychiatry 1995; 152:929-931.

Introduction:

^-           High frequency of depression in the hospitalized medically ill elderly (up to 45%^{Koenig HG et al  Clin Geriatr Med 1992;8:235-51})compromising long-term outcome.  Conventional ATD have delayed onset of action + either poor side effect profile (TCAs) or potential risk for pharmacokinetic interaction (SSRIs) ^{Frye C. American Journal of Health-System Pharmacy. 1997;54(21):2510-1}

-          Conclusions of  “Stimulants in the Treatment of Depression: A Critical Overview” : a) 1 placebo-controlled study out of 10 indicates a SS advantage of the psychostimulant, b) could be an interesting adjunct therapy to promote earlier onset of response in combination with conventional ATD and c) its suggested indication in medically ill elderly should be pursued with placebo-controlled trials. ^{Satel S. J Clin Psychiatry. 1989;50:241-9}

Study:

• 8-day, prospective, randomized, double-blind placebo-controlled, crossover design
• DSM IIIR diagnosis of major depressive disorder + medical illness such that inclusion in standard antidepressant trial would be impractical
• HDRS & MMS @ baseline, D4, D8 + 6W f/u
• Ritalin: 5 mg po qam/q noon x 2 d, then 10 mg x 2 d
• 16 patients (11W/5M) mean age 72.3, however 3 Pt dropped out < D4 => 10 started c active drug while 3 c placebo!
• RM ANCOVA, Rx W-S factor, order B-S factor and baseline covariate

Results:

• For HDRS => SS Rx effect & Rx by Order interaction.
  For the Ritalin-placebo group (n=10): differences between baseline-Ritalin, baseline-placebo but not Ritalin-placebo
  For the placebo-Ritalin group (n=3): difference between baseline-Ritalin
• (No difference between both groups on baseline, D4 & D8)
• No difference for MMS
• 7/13 had >55% HDRS decrease, 3/13 30-55% & 3/13 < 30%
• F/u (open design) for 9/13, 5 of those still on Ritalin maintained >= 55% HDRS reduction while the other 4 returned within 35% of baseline

Comments:

• Appears to be the only study built to address the efficacy of monotherapy with Ritalin in elderly medically ill depressed Pts
• Small imbalanced sample
• Inadequate use of HDRS
• No effect size calculable
• Not built to appropriately investigate outcome @ 6W
• Other trials have gone up to 60 mg/d

Bottomline:

1. Poorly convincing evidence suggesting rapid significant antidepressant effect of Ritalin within a small sample of elderly medically ill depressed patients
2. No evidence of sustained effect
3. Study does not address distinction between fatigue & depression
   => In study investigating fatigue within 109 non-MDD HIV Pts (excluding 35 dropouts), methylphenidate (max 60 mg/d) & pemoline (max 150 mg) showed SS greater improvement vs. placebo. Of note the difference was greater for affective subscale of Piper Fatigue Scale vs. sensory or energy subscales).  While Beck Depression Inventory SS correlated c total PFS, the correlation was stronger for severity & cognitive subscales vs. affective or sensory ones suggesting that the antidepressant effects reflect positive impact of decreased fatigue rather than direct antidepressant effect. ^{Breibart W et al. Archives of Internal Medicine. 2001;161(3):411-20}

Additional comments:

In a n=2 case report, 10 mg/d of methylphenidate prescribed to 76 and 54 y/o men did not accumulate despite chronic renal failure treated with either hemodialysis or peritoneal dialysis. ^{Stiebel V. Psychosomatics. 1994;35(5):498-500}

Methylphenidate acts as a reuptake inhibitor while d-amphetamine stimulates the cytoplasmic release. ^{Little KY. J. Clin Psychiatry. 1993;54(9):349-55}

 

 

Return to top of the page

 

---