Table 1.
Summary of the 33 articles included in this review.
| Reference | Population | N | Treatment | Duration (weeks) | Measures | Results |
|---|---|---|---|---|---|---|
| Double-Blind Randomized Controlled Trials (DBRCTs) | ||||||
| Planansky and Heilizer, 1959b | SCZ (male)a | 59 | CPZ (N = 33) | 12 | Jump reaction time; | - CPZ: Jump Time Consistency (r = 0.46; p = 0.05) and A cluster, increased activity (r = 0.70; p = 0.01). |
| PBO (N = 26) | A/F clusters (Lorr Scale) | |||||
| - PBO: No significant associations. | ||||||
| Holden and Holden, 1970b | SCZ (male)a | 22 | CDX | 8 | 7-point global scale | - % WG significantly correlated with decreased associative and perceptual disturbance, motor activity, depression, emotional liability and confusion (r = 0.43; p < 0.01, R2 = 0.12) |
| TDZ | BPRS | |||||
| PBO | MIPRS | |||||
| - CDX: % WG significantly correlated with increased drive | ||||||
| - TDZ: % WG significantly correlated with decreased associative disturbance, agitation, drive and confusion | ||||||
| - CDX + TDZ: % WG significantly correlated with decreased confusion, thinking disturbance and paranoid reaction | ||||||
| - PBO: % WG significantly correlated with decreased associative and perceptual disturbance, motor activity, apathy, confusion, thinking disturbance and paranoid reaction (r = 0.47; p < 0.01, R2 = 0.16) | ||||||
| Poyurovsky et al., 2002 | SCZa | 30 | OLZ+PBO (N = 15) | 8 | SANS | - OLZ+PBO: No significant association with pos. (r = 0.45), neg. (r = 0.17) or disorganize (r = 0.20) symptoms |
| OLZ+FLX (N = 15) | SAPS | |||||
| - OLZ+FLX: No significant association with pos. (r = 0.08), neg. (r = 0.21) or disorganize (r = 0.11) symptoms | ||||||
| PROSPECTIVE STUDIES | ||||||
| Leadbetter et al., 1992b | SCZ | 21 | CLZ | 16 | BPRS | - Significant association of BRPS score with WG (r = 0.54, p = 0.01) |
| SAD | ||||||
| - TB difference between marked WG (≥10% increase) and non-marked WG (10.2 BPRS total points, p < 0.03; neg. symptoms, 3.5 points, p < 0.02). | ||||||
| Jalenques et al., 1996 | SCZ | 15 | CLZ | 84 | BPRS | - All patients showed same improvement at 20 weeks. |
| - Group 1: Achieved further improvement (maximally, 8% further decrease in BPRS) and had WG: (12.4 kg; p < 0.001 | ||||||
| - Group 2: Worsened (at worst BPRS of 89% initial improvement) and had NS WG on average. | ||||||
| Hummer et al., 1995 | SCZ | 81 | CLZ | NR | CGI | - No correlation between CS-WG (≥10% increase) and TB (statistics NR). |
| Bai et al., 1999b | SCZ | 96 | CLZ | 56 ± 24.8 | CGI | - Females (N = 46): Responders vs. non-responders: 10.9 kg vs. 4.3kg; p = 0.002. |
| - Males (N = 50): Responders vs. non-responders: 7.10 kg vs. 7.10 kg; p = NS | ||||||
| Meltzer et al., 2003b | SCZ | 74 | CLZ | 24 | BPRS | - Controlled for age, baseline psychopathology and baseline weight (no effect of sex) |
| SAD | SADS-C | |||||
| - At 6 weeks, % WG correlated with total BPRS (β = 0.28, p = 0.05), BPRS pos. symptoms (β = 0.17, p = 0.01), SADS-C severity of delusions (β = 0.03, p = 0.05), SAPS total (β = 0.10, p = 0.04) | ||||||
| GAFS | ||||||
| SANS | ||||||
| SAPS | ||||||
| QLS | - At 24 weeks, % WG correlated with total BPRS (β = 0.20, p = 0.01), and SANS (β = 0.13, p = 0.002) | |||||
| Garyfallos et al., 2003b | SCZ | 50 | OLZ (N = 25) | 8 | PANSS | - OLZ: WG correlated with BMI (r = 0.69, p < 0.01), triglyceride levels (r = 0.71, p < 0.001) and total cholesterol (r = 0.36, p > 0.05) |
| SAD | RIS (N = 25) | |||||
| SPFa | ||||||
| - RIS: WG not correlated with BMI (p > 0.10), triglycerides (p > 0.10), or total cholesterol (p > 0.05). | ||||||
| Lane et al., 2003b | SCZa | 146 | RIS | 6 | PANSS | - Controlled for treatment duration, age, sex, type of SCZ, dose and baseline weight |
| NOSIE | ||||||
| - WG correlated with pos. symptoms (β = 0.06, p = 0.009), neg. symptoms (β = 0.08, p = 0.0005), cognitive functions (β = 0.08, p = 0.003) and NOSIE (β = 0.04, p < 0.0002) | ||||||
| - Significant difference in WG between responders and non-responders (difference = 0.51kg, p = 0.007). | ||||||
| Konarzewska et al., 2013b | SCZ (male) | 31 | OLZ (N = 14) | 8 | PANSS | - OLZ: PANSS score not associated with BMI (r = 0.03, p = 0.91) |
| RIS (N = 17) | ||||||
| - TB was significantly correlated with increased insulin levels for PANSS-Total, pos., and general symptoms, but not neg. symptoms (N = 10). | ||||||
| - RIS: PANSS score was not associated with BMI (r = 0.36, p = 0.15); | ||||||
| - TB was not correlated with insulin levels (N = 16) | ||||||
| Zhang et al., 2003b | SCZ | 117 | CPZ (N = 66) | 10 | PANSS | - No significant association between change in BMI and TB (p > 0.05), after controlling for age, sex, illness duration, baseline BMI, baseline psychopathology, dose and DRD2 genotype |
| RIS (N = 43) | ||||||
| CLZ (N = 4) | ||||||
| Other (N = 4) | ||||||
| Zhang et al., 2004b | SCZ | 46 | CPZ | 10 | PANSS | - No significant correlation of TB with BMI, AIWG, waist-to-hip ratio, or MRI-measures subcutaneous/ intra-abdominal fat |
| RIS | ||||||
| Venkatasubramanian et al., 2010b | SCZ | 27 | OLZ (N = 12) | 12 | SANS | - SANS score significantly correlated with BMI (r = 0.40, p = 0.04) and leptin levels (r = 0.46, p = 0.02). |
| RIS (N = 12) | SAPS | |||||
| FPX (N = 3) | ||||||
| Schwarz et al., 2012b | SCZa | 77 | VAR | 6 - 156 | PANSS | - Early relapsers (< 33 weeks; N = 9) vs. late relapsers (N = 9): |
| - had lower BMI (4 kg/m2, p = 0.04), leptin (change ratio = 0.18, p = 0.03) proinsulin (change ratio = 0.29; p = 0.02) and higher TGF-α (change ratio = 1.85; p = 0.018) | ||||||
| - Lower leptin (Δratio = 0.21; p = 0.024) and insulin (Δratio = 0.50; p = 0.024) among other molecules but not BMI at time of relapse. | ||||||
| - Had 1.66 kg/m2 lower BMIΔ and less likely to have an increase in leptin, insulin and c-peptide in first 6 weeks. | ||||||
| CHART REVIEW | ||||||
| Planansky, 1958b | SCZ | 123 | CPZ | 20 | Nurse evaluation | - 13.8% WG with marked improvement, 11.1% WG with minimal improvement and 5.9% WG with no improvement. (p < 0.001) |
| Lamberti et al., 1992b | SCZ | 36 | CLZ | 24 | BPRS | - r = 0.31; p < 0.1 but NS. |
| Umbricht et al., 1994 | SCZ | 84 | CLZ | 24 | BPRS, | - At 6 weeks: r = 0.10; p = NS, Responders (N = 12) did not have significantly more WG than nonresponders (N = 44). |
| CGI | ||||||
| - At 12 weeks (N = 26): r = 0.28; p = NS, Responders (N = 10) did not have significantly more WG than nonresponders (N = 16). | ||||||
| - At 24 weeks: Responders (N = 19) did not have a higher %WG or a higher cumulative incidence of 5, 10, or 20% WG than nonresponders (N = 20) did in first 8 weeks. | ||||||
| Bai, 2006 | SCZ (Taiwanese) | 140 | CLZ | 384 | CGI | - Controlling for baseline BMI, gender, age, dose, use of other psychotropic drugs: β = 0.297; p = 0.007 at 24 weeks (N = 85). β = 0.274; p = 0.03 at 384 weeks (N = 55). |
| - Early responders (CGI 1-2 after 14 months) vs. early nonresponders: 13.8 (8.4) kg vs. 4.5 (12) kg; p = NR. | ||||||
| - No sex differences after controlling for baseline BMI. | ||||||
| Hung et al., 2010b | SCZ (Taiwanese) | 97 | RIS | 4 | MSPI | - Female patients (N = 58): β = 3.26; p < 0.01. Adjusting for baseline BMI and MSPI: β = 2.51; p < 0.01. Male patients (N = 39): β = 0.11; p > 0.05. Adjusting for baseline BMI and MSPI: β = 0.30; p > 0.05. |
| Sharma et al., 2011 | SCZ | 46 | RIS (41.8%), | 30.14 (18.03) | CGI | •[-]r = 0.3; p = 0.04 |
| OLZ (27.9%), | ||||||
| BAD | CLZ (14%), Others | |||||
| POST-HOC | ||||||
| Agid et al., 2013b | SCZ, SAD | 94 | OLZ | 24 | GAF | - patients with CS-WG at week 6 had significantly less improvement in GAF scores at week 24 (p = 0.031). |
| ZIP | ||||||
| - Among patients with CS-WG, mean WG was lower in completers compare to those who dropped out prior to 24 weeks (p = 0.002). | ||||||
| Bustillo et al., 1996 | SCZ | 39 | CLZ (N = 20) | 10 | BPRS, SANS | - r = 0.48; p = 0.002 (%BPRS,%WG). No significant correlations with SANS total score or BPRS subscores. |
| HAL(N = 19) | ||||||
| - CLZ: r = 0.57; p = 0.01 (%BPRS,%WG). No significant correlation found in 48 week open CLZ trial (N = 33) follow-up to original study | ||||||
| - HAL: r = 0.19; p = 0.40 (%BPRS,%WG) | ||||||
| Basson et al., 2001 (Study 1)1 | SCZ, SAD, SPF | 1996 | OLZ (N = 1336), | 6 | BPRS | - After adjusting for age, dose, gender, baseline BMI, ethnicity, appetite and treatment group: WG difference between responders (BPRS ≤ 18) and nonresponders: ~0.5 kg (p = 0.003). |
| HAL (N = 660) | ||||||
| Basson et al., 2001 (Study 2) | SCZ, SAD, SPF | 339 | OLZ (N = 172), RIS (N = 167) | 6 | BPRS | - After adjusting for age, dose, gender, baseline BMI, ethnicity, appetite and treatment group: WG difference between responders (BPRS ≤ 17) and nonresponders: ~1.5 kg (p = 0.001). |
| Czobor et al., 20022b | SCZ, SAD | 151 | - OLZ (N = 38) | 14 | PANSS | - OLZ: Controlling for baseline psychopathology and baseline body weight: partial r = 0.57; p < 0.0003 (total), partial r = 0.61; p = 0.0001 (general psychopathology), partial r = 0.55; p = 0.0001 (pos. symptoms), partial r = 0.34; p = 0.042 (neg. symptoms). WG difference between responders and deteriorators: 9 kg (p = 0.005). 55% of responders had marked weight gain (>10%) vs. 0% of deteriorators (p < 0.001) |
| CLZ (N = 38) | ||||||
| RIS (N = 39) | ||||||
| HAL (N = 36) | ||||||
| - CLZ: Controlling for baseline psychopathology and baseline body weight: partial r = 0.39 p < 0.02 (total), partial r = 0.39; p = 0.018 (PANSS general psychopathology), partial r = 0.31; p = 0.07 (PANSS pos. symptoms), and no significant correlation for PANSS neg. symptoms subscale. WG difference between responders and deteriorators: 5 kg (p = 0.002). 40% of responders had marked weight gain (>10%) vs. 0% of deteriorators (p = 0.02). | ||||||
| - RIS: partial r = 0.00; p = NS. Responders did not gain significantly more than deteriorators. | ||||||
| - HAL: partial r = 0.30; p < 0.06. Responders did not gain significantly more than deteriorators. | ||||||
| Ascher-Svanum et al., 2005a | SCZ | 248 | OLZ (N = 187) | 6 | BPRS | - OLZ: r = 0.43; p < 0.001; slope = 0.107 kg/unit. Responders had significantly greater WG than deteriorators: 4.71 kg vs. 1.17 kg; p < 0.001. Adjusting for treatment duration:partial r = 0.21; p = 0.003; slope = 0.064 kg/unit, 4.03 kg vs. 2.24 kg; p = 0.029. |
| PBO (N = 61) | ||||||
| - PBO: r = 0.36; p = 0.004; slope = 0.052 kg/unit, 0.40 kg vs.−1.83 kg; p = 0.003. Adjusting for treatment duration: partial r = 0.41; p = 0.001; slope = 0.072 kg/unit, 0.80 kg vs.−2.29 kg; p < 0.001. | ||||||
| Ascher-Svanum et al., 2005b1 | SCZ, SAD, SPF | 1996 | OLZ (N = 1336) | 6 | PANSS,BPRS, MADRS, SF-36 | - OLZ, HAL: Controlling for group, gender, group-gender interaction, baseline weight, baseline psychopathology and treatment duration: BPRS pos. and neg. symptoms: B = 0.038; p < 0.001. Depressive symptoms (MADRS, BRPS): B = 0.030; p < 0.001. Mental functioning (SF-36): B = 0.026; p = 0.047. Physical functioning (SF-36): B = 0.028; p = 0.021). No gender effect. |
| HAL (N = 660) | ||||||
| - OLZ: Responders vs. deteriorators (BPRS): 2.49 kg vs. 1.42 kg; p < 0.001 (BPRS), 2.37 kg vs. 0.59 kg; p < 0.001 (PANSS). Controlling for baseline weight, treatment duration and baseline psychopathology: partial r = 0.15; p < 0.001 (PANSS). Controlling for baseline weight and baseline psychopathology: partial r = 0.24; p < 0.001 (PANSS) | ||||||
| - HAL: Responders vs. Deteriorators: 0.08 kg vs.−0.44 kg; p = 0.087 (BPRS), 0.15 kg vs.−0.55 kg; p = 0.01 (PANSS). Controlling for baseline weight, treatment duration and baseline psychopathology: partial r = 0.11; p = 0.006 (PANSS). Controlling for baseline weight and baseline psychopathology: partial r = 0.10; p = 0.013. | ||||||
| Kinon et al., 20051 | SCZ, SAD, SPF | 1191 | OLZ | 6 | BPRS | - CS-WG (N = 183) vs. nonCS-WG (N = 1008): 17.1 (11.1) vs. 21.0 (12.5); p = 0.0009 (BPRS at 6 weeks), 54% vs. 45%, p = 0.03 (% of BPRS < 18 at 6 weeks) |
| Müller et al., 20052b | SCZ, SAD | 59 | CLZ, HAL, OLZ, RIS | 14 | PANSS | •[-]ρ = 0.51, p < 0.001. |
| Zipursky et al., 2005 | SCZ, SAD, SPFa | 263 | OLZ (N = 131) | 96 | PANSS | - r = 0.21; p = 0.02. Correlation only seen in week 1 and 6 for both treatment groups (week 12 onward the correlation was NS). |
| HAL (N = 132) | ||||||
| Procyshyn et al., 2007 | SCZ | 55 | RIS, PBO | 8 | PANSS | - WG not associated with improvement of PANSS total, pos. or neg. subscores (p ≥ 0.1). Controlling for WG: TG (mmol/L): B = 3.74, β = 0.285, p < 0.037 (total), B = 1.565, β = 0.325; p = 0.017 (neg. symptoms). TC: B = 2.36, β = 0.370; p = 0.007 (neg. symptoms). TG, TC did not correlate with pos. symptoms (p > 0.25). Responders vs. nonresponders had significantly greater increases in TG (0.8 mmol/L; p = 0.004) and TC (0.52 mmol/L; p = 0.008) but not weight (p = 0.917). |
| Hermes et al., 2011b | SCZ | 865 (1245 mixed model) | OLZ, | 72 | PANSS | - CS-WG vs. nonCS-WG: decrease of 14.1 (15.7) points vs. 5.6 (16.2) points; p = NR. Controlling for baseline psychopathology, baseline BMI, treatment group, age, illness duration and investigator site: 12 weeks: partial R2 = 0.01; p < 0.001; slope = 0.28 (points/%BMI), 72 weeks: slope = 0.21; p < 0.001. No significant differences were found between treatment groups in estimating correlation between PANSS and BMI. No association found between PANSS and total cholesterol or triglycerides. |
| RIS, | ||||||
| QTP, | ||||||
| ZIP, | ||||||
| PPZ | ||||||
| - OLZ: 12 weeks: partial R2 = 0.02; p = 0.010; slope = 0.37, 72 weeks: slope estimate = 0.17; p = 0.001 | ||||||
| - RIS: 12 weeks: NS, 72 weeks: slope = 0.24; p = 0.002 | ||||||
| - QTP: 12 weeks: NS, 72 weeks: slope = 0.18; p = 0.017 | ||||||
| - ZIP: 12 weeks: NS, 72 weeks: slope = 0.31; p = 0.005 | ||||||
| - PPZ: 12 weeks: partial R2 = 0.02; p = 0.022; slope = 0.56, 72 weeks: NS | ||||||
| Kemp et al., 2013 | SCZa | 107 | OLZ (N = 72) | 6 | BPRS | - r = 0.31; p < 0.01 (%BPRS,%WG), r = 0.29; p = 0.015 (%BPRS/BMI z-score). Controlling for age, gender, race, age at illness onset, BMI z-scoreΔ, baseline BMI z-score, blood pressure, HDL, LDL, BPRS, and CGI: only BMI z-score predicted WG (p = 0.01). Became NS when controlling for duration of treatment (p = 0.12). TB difference between CS-WG and nonCS-WG: 13.7%; p = 0.04. CS-WG did not predict higher rates of response (≥ 30% improvement in BPRS), early response (≥20% at 2 weeks) or remission. |
| PBO (N = 35) | ||||||
| - PBO: r = 0.31; p = 0.08 (%BPRS/%WG), r = 0.33; p = 0.06 (%BPRS/BMI z-score). TB difference CS-WG vs. nonCS-WG: 17.3%; p = 0.36 | ||||||
a
Population was completely or partially antipsychotic naïve.
b
Study took measures to ensure compliance.
1
All analyzed (Tollefson et al., 1997).
2
Both analyzed (Volavka et al., 2002).
Bolded trials do not show evidence for an AIWG-TB association. The bolded and italicized trial shows evidence for an inverse AIWG-TB association.
All correlations where therapeutic benefit (TB) increases with WG/metabolic changes are shown as positive for simplicity. If not specified, correlation is between Δ total scores and absolute WG (kg). Response is ≥ 20% decrease in score unless otherwise specified, nonresponse is anything less than this and deterioration is any increase in psychopathology. CS-WG is ≥ 7% BMI/body weight, unless otherwise specified. Absolute WGs and scores represent averages (1SD). Significance is p ≤ 0.05. APs separated by commas were analyzed together.
Pos, positive; neg, negative; NR, Not reported; NS, nonsignificant; WG, weight gain; CS-WG, clinically significantly weight gain; TB, therapeutic benefit; TG, triglyceride; TC, total cholesterol; SCZ, schizophrenia; SAD, schizoaffective disorder; SPF, schizophreniform disorder; BAD, bipolar disorder; AP, antipsychotics; AAP, atypical antipsychotics; TAP, typical antipsychotics; CPZ, chlorpromazine; PBO, placebo; OLZ, olanzapine; CLZ, clozapine; RIS, risperidone; HAL, haloperidol; QTP, quetiapine; ZIP, ziprasidone; PPZ, perphenazine; FPX, flupenthixol; CDX, chlordiazepoxide; TDZ, thioridazine; FLX, fluoxetine; VAR, variety of antipsychotics; SANS, Scale for Assessment of Negative Symptoms; SAPS, Scale for Assessment of Positive Symptoms.