- Email: [email protected]
Unexplained transient hyperphosphatemia is more common in acute psychiatric disorders than in acute medical–surgical conditions
Psychiatry Research 117 (2003) 237–243
Unexplained transient hyperphosphatemia is more common in acute psychiatric disorders than in acute medical–surgical conditions Ali Ahmadi, Edward C. Lauterbach*, Nasreen Malik, John Gonzales Department of Psychiatry and Behavioral Sciences, Mercer University School of Medicine, 655 First Street, Macon, Georgia 31201, USA Received 25 January 2001; received in revised form 18 October 2002; accepted 22 January 2003
Abstract Increased serum inorganic phosphorus associated with elevated serum calcium has been demonstrated to coincide with the onset of agitation and mania in periodic psychoses and bipolar disorders. We tested the hypothesis that unexplained transient hyperphosphatemia (UTHP) is more common in patients with psychiatric disorders than in controls with medical or surgical conditions. We studied 100 patients admitted to a psychiatric ward and 100 controls admitted to a medical–surgical ward. All subjects (patients and controls) underwent acute admission to the same general hospital. The serum phosphorus was measured upon admission and, if elevated, followed during the hospital course. Twenty patients (20%) with psychiatric disorders had unexplained hyperphosphatemia compared with four medical–surgical controls (4%). UTHP occurred in six patients with psychiatric disorders and no controls. Hypophosphatemia did not occur in subjects with psychiatric disorders. This study shows an increased incidence of UTHP in acutely ill, hospitalized patients with psychiatric disorders relative to acutely ill, hospitalized controls with medical–surgical conditions. These data extend previous findings by linking UTHP to acute psychiatric disturbances across varied psychiatric diagnoses independent of hypercalcemia. Potential explanations include trazodone administration and transient hypocalcemia. 䊚 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Phosphorus; Major depression; Schizophrenia; Bipolar disorder; Alcohol disorders; Substance abuse disorders
1. Introduction Elevated serum phosphorus levels accompanied by increased serum calcium have been observed *Corresponding author. Tel.: q1-473-301-2107; fax: q1478-301-5337. E-mail address: [email protected] (E.C. Lauterbach).
to coincide with exacerbations of rapid-cycling psychotic disorders and mania (Carman and Wyatt, 1979b). Additional studies have shown a significant reduction in agitation ratings in patients with agitated psychotic disorders following administration of probenecid (a potent phosphaturic agent) and calcitonin (Carman and Wyatt, 1979a,c). These increases in serum calcium and phosphorus, which precede behavioral changes, have been sug-
0165-1781/03/$ - see front matter 䊚 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0165-1781(03)00021-0
238
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
gested as a cause rather than an effect of psychotic agitation and mania (Snowdon et al., 1976). A significant number of patients with psychiatric disorders have disturbed behavior upon admission. We were struck by a clinical impression of remarkably frequent serum phosphorus elevations without calcium perturbations on our psychiatric ward and so we decided to study this phenomenon. Although previous studies evidenced a linkage between hyperphosphatemia and agitation associated with hypercalcemia observed in mania or psychosis (Snowdon et al., 1976; Carman and Wyatt, 1979a,b,c), medical–surgical controls were not employed. The aim of our study was to compare the serum phosphorus level of patients with psychiatric disorders to that of controls with medical or surgical conditions upon admission to the same hospital. We tested the hypothesis that unexplained transient hyperphosphatemia (UTHP) is more common in patients with psychiatric disorders than in control patients with medical or surgical conditions. 2. Methods We prospectively examined laboratory and clinical data on 100 consecutively admitted patients with psychiatric disorders and compared them to 100 consecutively admitted patients with medical– surgical conditions. All patients (psychiatric and medical–surgical) were admitted during the same time period, and patients with psychiatric disorders were admitted to a general psychiatric ward while patients with medical–surgical conditions were admitted to a medical–surgical ward. Both wards were part of a 519-bed community teaching hospital. Phosphorus levels for each group were determined on admission using conventional clinical laboratory techniques (Dryer and Routh, 1963) according to norms determined by the hospital clinical pathology department. Data collected for each subject included age, gender, psychiatric diagnoses, medicalysurgical diagnoses, all medications, their doses, routes and dosing intervals, phosphorus (normative reference level 2.4–4.6 mgydl), calcium, magnesium, blood urea nitrogen (BUN), creatinine, BUNycreatinine ratio, T3 uptake, T4, thyroid index, TSH and a
Table 1 Hyperphosphatemic etiologies Serum protein binding (lymphocyte dyscrasias) Decreased renal excretion Renal insufficiency Hypoparathyroidism Pseudohypoparathyroidism (Types I and II) Tumoral calcinosis Pseudoxanthoma elasticum Infantile hypophosphatasia Hyperostosis Hyperthyroidism Growth hormone Adrenal insufficiency Biphosphonate therapy Increased intestinal absorption Phosphorus-containing cathartics Medication with vitamin D compounds Granulomatous diseases producing vitamin D (sarcoidosis, tuberculosis) Internal redistribution Acute metabolic acidosis Acute respiratory acidosis Reduced insulin level Clonidine administration Cellular release Rhabdomyolysis Organ infarction Tumor lysis (lymphoma, metastatic small cell carcinoma) Thyrotoxicosis Acute hemolysis Parenteral administration Intravenous phosphate salts Phospholipid infusion Medical causes of hyperphosphatemia.
listing of all other abnormal laboratory and ancillary study findings. Parathormone levels were not obtained. Clinical charts were intensively reviewed for evidence of medical causes of phosphorus disturbances (Bennett and Plum, 1996). Hyperphosphatemia was considered to be unexplained if the chart review disclosed no evidence suggesting any of the possible etiologies of an elevated serum phosphorus level. A list of causes of hyperphosphatemia is provided in Table 1. We screened clinical data conservatively, construing even minor abnormalities on clinical exam or laboratory data to be suggestive of an explainable cause (e.g.
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
mildly elevated creatinine consistent with renal insufficiency). We assessed causation in all cases of disturbed phosphorus regardless of whether the abnormal value was transiently or persistently abnormal. Transient hyperphosphatemia was defined by the return of the phosphorus level back to the normal range on repeated determinations within 120 h of admission (evident on 2–5 repeated determinations), whereas persistent hyperphosphatemia was defined by consistently elevated values on repetitive determinations. Unexplained transient hyperphosphatemia (UTHP) was defined by absence of an explainable cause and transience, according to the above criteria. 2.1. Hypothesis testing Frequencies of hypophosphatemia, hyperphosphatemia, and UTHP were determined for the psychiatric and medical surgical groups. We hypothesized that UTHP is more common in patients with psychiatric disorders than in patients with medical–surgical conditions. We tested the hypothesis by comparing UTHP rates between the two groups using the one-tailed Pearson’s x2 with Yates correction factor. We assessed comparability of the two groups using a two-tailed two-way analysis of covariance of group and gender with age as a covariate. In the event of significance, a two-tailed two-way analysis of covariance of phosphorus by group with the significant findings as the covariate was determined. 2.2. Additional data exploration We subsequently explored the data for factors that might distinguish patients with UTHP from other patients by comparing patients with psychiatric disorders with UTHP to two additional sets of controls: (1) subjects with psychiatric disorders lacking UTHP; (2) subjects with psychiatric disorders lacking phosphorus aberration of any kind (i.e. lacking hyperphosphatemia, since hypophosphatemia was not observed in subjects with psychiatric disorders). We compared the subjects with UTHP to each of the two psychiatric control groups for demographic and clinical variables using the two-tailed Student’s t-test for analysis of
239
age and two-tailed Fisher’s exact test for categorical variables. 3. Results Initial inspection of laboratory values revealed hyperphosphatemic values in 22 psychiatric (ns 100) and 8 medical surgical (ns100) subjects. Means and standard deviations of phosphorus values for the groups were: UTHP (ns6 psychiatric subjects) 5.35"S.D. 0.3; No UTHP (ns94 psychiatric subjects) 4.01"0.81; Normal Phosphorus (ns76 psychiatric subjects) 3.84"0.50; Medical– Surgical Subjects (ns100) 3.43"0.66. Intensive chart review for causes of phosphorus disturbance revealed potential medical explanations in two psychiatric (hemolytic anemia, dialysis) and four medical–surgical (respiratory acidosis in two, diabetes mellitus, renal insufficiency) hyperphosphatemic cases. Hypophosphatemic values occurred in 0 psychiatric and 10 medical–surgical subjects, consistent with the somewhat lower mean phosphorus level in the medical–surgical group. Nine cases of hypophosphatemia in medical–surgical subjects revealed potential medical exlanations in each, including diuretic administration in four, malabsorption in two, alcoholism in one, insulin therapy in one, and depression (which was diagnosed as complicating the primary medical illness) in one. Thus, unexplained hyperphosphatemia (including UTHP cases) occurred in 20 psychiatric and four medical–surgical cases whereas unexplained hypophosphatemia occurred in no psychiatric and one medical–surgical subject (Table 2). Subjects with UTHP are displayed in Table 3. 3.1. Hypothesis testing UTHP was significantly more common in subjects with psychiatric disorders (Table 2) than in subjects with medical–surgical disease (6 of 100 psychiatric cases vs. 0 of 100 controls, one-tailed Yates-corrected Pearson’s x2 s4.2955, d.f.s1, P0.02). Subjects with psychiatric disorders (mean age 41.12"S.D. 13.65 years) were significantly younger than subjects with medical–surgical conditions (59.95"17.41, ANOVA Fs68.591, d.f.s 1, P-0.000000001). There was no significant
240
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
Table 2 Comparison of subjects with psychiatric disorders to medical–surgical controls Condition
Psychiatric subjects (ns100)
Medical–surgical controls (ns100)
Hyperphosphatemia Unexplained Transient (UTHP)* Hypophosphatemia Unexplained Transient Demographics Men** Women** Age†
22 20 6 0 0 0
8 4 0 10 1 0
50 50 40.64"S.D.13.49
42 58 60.03"S.D.16.92
*UTHP was significantly more common in subjects with psychiatric disorders than in controls (one-tailed Yates-corrected Pearson’s x2s4.2955, d.f.s1, P-0.02). **Gender difference was not significant between groups (Pearson’s x2 s1.284, d.f.s1, Ps 0.26). †Subjects with psychiatric disorders were significantly younger than controls (ANOVA Fs79.284, d.f.s1, Ps 0.0000000001).
difference in gender between the two groups (psychiatric: 50 males, 50 females; medical–surgical: 42 males, 58 females, Fs0.110, d.f.s1, Ps0.74). The analysis of covariance of phosphorus level by group with age as a covariate disclosed a significant effect of group (4.13"0.98 vs. 3.43"0.66, Fs17.137, Ps0.00005) but not of age (Fs2.713, d.f.s1, Ps0.1) on phosphorus level.
transient hypocalcemia (Ps0.09) compared to the ‘Normal Phosphorus’ group. In summary, UTHP was significantly more common in psychiatric patients than in medical-surgical patients and did not relate to other clinical variables, suggesting the possibility that UTHP may be linked to acute psychiatric conditions. 4. Discussion
3.2. Additional data exploration We explored the results for variables differing between subjects with UTHP and the two psychiatric control groups because larger samples in future studies might discriminate UTHP clinical features detectable only as non-significant trends in this study. Comparisons of the six subjects with UTHP to the two psychiatric control groups are displayed in Table 4. There were no significant differences (Bonferroni-corrected as0.0016) between subjects with UTHP and either control group for any of the demographic or clinical comparison variables (listed in Table 4). Without Bonferroni correction, there were non-significant trends toward younger age (Ps0.09) and more trazodone administration (Ps0.06) in the UTHP group compared to the ‘No UTHP’ group, and significantly more trazodone administration (Ps 0.04) and a non-significant trend toward more
This is the first report of a prospective controlled study demonstrating UTHP in acute, hospitalized subjects with psychiatric disorders when compared with acute, hospitalized subjects with medical– surgical disease. The significant difference in age between the two groups reflects the difference in age-related risks for psychiatric and medical–surgical conditions. Transient increases in serum phosphorus coinciding with elevated serum calcium and rapid cycling of psychotic agitation and mania were demonstrated in the past (Snowdon et al., 1976; Carman and Wyatt, 1979a,b,c). The aim of our study was to ascertain prospectively the serum phosphorus level of 100 psychiatric patients upon admission to the psychiatric ward, follow them up if levels were elevated during the hospital course, and compare them in the same fashion with 100 patients admitted to the medical–surgical ward. Previous studies did not use such controls and
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
241
Table 3 Characteristics of patients with psychiatric disorders and UTHP Case
Age
Sex
Diagnosis
Medications
1
41
F
Major depression
2
38
M
Cocaine-induced psychotic disorder
3
22
M
Substance-induced psychotic disorder
4
42
M
Schizophrenia chronic paranoid type
5
31
F
6
32
M
Alcohol-induced mood and psychotic disorders Adjustment disorder with depressed mood Schizoaffective disorder
Bupropion 100 mg Nitroglycerine Risperidone 6 mg Fluoxetine 60 mg Trazodone 100 mg Risperidone 4 mg Valproic acid 1000 mg Trazodone 100 mg Fluphenazine decanoate 50 mgymo Benztropine 2 mg Trifluoperazine 100 mg Ibuprofen 600 mg
Quetiapine 800 mg Venlafaxine 75 mg Diphenhydramine 100 mg Trazodone 200 mg Benztropine 2 mg Haloperidol 15 mg Valproic acid 1000 mg
The age, gender, diagnosis at discharge, and medication regimen are displayed for each subject with UTHP.
Table 4 Comparison of subjects with UTHP to other subjects with psychiatric disorders
Male Female Age Psychotic diagnosis Psychotic condition Mood diagnosis Mood condition Antidepressants Antipsychotic Valproic acid Benztropine Trazodone Lithium Substance abuse Transient hypocalcemia Elevated T3 uptake test
UTHP (ns6)
No UTHP (ns94)
Normal phosphorus (ns78)
4 2 34.33qS.D.7.55 1 4 2 3 4 4 2 2 3 0 3 3 3
46 48 40.95qS.D.13.67† 30 38 44 48 35 41 16 9 15† 6 28 15 24
36 42 42.11qS.D.13.88 24 31 36 38 28 33 14 7 10* 5 23 15† 21
Comparison of 6 subjects with psychiatric disorders and UTHP to the 94 subjects with psychiatric disorders lacking UTHP (‘No UTHP’) and the 78 subjects with psychiatric disorders lacking phosphorus aberrations of any kind (‘Normal Phosphorus’). †P0.1, *P-0.05.
242
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
their main objective was to demonstrate transient hypercalcemia as a trigger of the agitated state observed in rapidly cycling acute exacerbations of mania or psychosis. In contrast with previously reported studies, we did not detect transient hypercalcemia in our patients with psychiatric disorders (Snowdon et al., 1976; Carman and Wyatt, 1979a,b,c). Rather, we identified hyperphosphatemia independent of hypercalcemia. Although our study relied on the prospective identification of UTHP in 100 patients with psychiatric disorders and 100 patients with medical– surgical illness, and used conventional standardized laboratory assessment, certain methodological limitations merit consideration. These limitations include the use of clinical diagnoses (as opposed to structured clinical diagnostic interviews), the possibility of undetected physical illness, age and medication differences between the groups, possible laboratory error, and insufficient power to detect differences between UTHP and non-UTHP psychiatric subgroups. The use of clinical rather than structured interview diagnoses, however, does not alter our UTHP finding, ensures comparability with controls (medical–surgical patients were also diagnosed clinically), and allows generalization of our findings to other clinical populations. Undetected physical illness in subjects with psychiatric disorders remains a possibility although this is less likely in light of the transiency of UTHP, younger age of these patients with psychiatric disorders, and reciprocal cross-training of each physician on both wards used in the study. We chose medical–surgical subjects from the same hospital as a comparison group because this group would be expected to have high rates of phosphorus abnormalities and constitute a rigorous test of the hypothesis, but we did not anticipate such a significant difference in subject’s age. Nevertheless, neither younger age nor the medications in our subjects with psychiatric disorders are known to increase adult phosphorus levels, and phosphorus levels were not significantly related to age in our subjects. Clinical laboratory false-positive rates for hyperphosphatemia would be expected to be identical for the two groups since levels were determined over the same period of time using the same assay. Moreover, false-positive hyperphos-
phatemia rates would be expected to be higher in the medical–surgical patients since these patients often undergo a greater number of laboratory tests, thus multiplying the chances of laboratory error and thereby biasing this study against finding an increase in subjects with psychiatric disorders. Finally, we did not originally design this study with the intent of looking for predictive differences between UTHP and non-UTHP psychiatric subgroups. Overall, the use of our medical–surgical control group with its possible bias toward increased UTHP makes it likely that the increased rate of UTHP in subjects with psychiatric disorders is a legitimate finding. There has been some previous speculation in the literature regarding ionic involvement in the pathophysiology of psychiatric illnesses. In one study, no difference in serum phosphorus level was found between patients with chronic schizophrenia and patients with non-psychotic mental retardation (Yassa et al., 1979). In another study, haloperidol treatment did not reduce the serum levels of phosphorus or calcium but reduced the magnesium level (Jabotinsky-Rubin et al., 1993). Interestingly, among 88 patients with end-stage renal disease, higher phosphorus levels were present in patients undergoing hemodialysis who were younger and more depressed in comparison to other hemodialysis patients (Kimmel et al., 1995). In the same study higher phosphorus levels correlated with aberrant behavior and poor dialysis compliance (Kimmel et al., 1995). Metabolic stress associated with acute psychiatric exacerbation or membrane phospholipid metabolic disturbances (Fukuzako et al., 1999) in our patients might possibly relate to transient hyperphosphatemia. A transient endocrinologic or metabolic disturbance is suggested by the over-representation of transient hypocalcemia in psychiatric subjects with UTHP relative to psychiatric subjects with normal phosphorus. It is also possible that trazodone administration may predispose to UTHP since trazodone administration was over-represented in UTHP subjects, although a MEDLINE literature search did not disclose an association. Alternatively, elevated phosphorus may reflect otherwise subclinical rhabdomyolysis, similar to
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
increased serum creatine phosphokinase seen in psychotic exacerbations. Although there was no evidence of clinical rhabdomyolysis in UTHP subjects, one had an elevated SGOT and another had an elevated LDH. In any event, these data extend previous findings by linking UTHP to acute psychiatric disturbances across various psychiatric diagnoses independent of persistent hypercalcemia. It will now be of interest to study patients compared to controls matched for age, gender, and extent of medical illness. References Bennett, J.C., Plum, F., 1996. Cecil Text Book of Medicine. 21st ed. WB Saunders Co, Philadelphia. Carman, J.S., Wyatt, R.J., 1979a. Calcium: bivalent cation in the bivalent psychoses. Biological Psychiatry 14, 295–336. Carman, J.S., Wyatt, R.J., 1979b. Calcium: pacesetting the periodic psychoses. American Journal of Psychiatry 136, 1035–1039.
243
Carman, J.S., Wyatt, R.J., 1979c. Use of calcitonin in psychotic agitation or mania. Archives of General Psychiatry 36, 72– 75. Dryer, R.L., Routh, J.I., 1963. Determination of serum inorganic phosphorus. Clinical Chemistry 4, 191–197. Fukuzako, H., Fukuzako, T., Kodama, S., Hashiguchi, T., Takigawa, M., Fujimoto, T., 1999. Haloperidol improves membrane phospholipid abnormalities in temporal lobes of schizophrenic patients. Neuropsychopharmacology 21, 542–549. Jabotinsky-Rubin, K., Durst, R., Levitan, L.A., Moscovich, D.G., Silver, H., Lerner, J., Van Praag, H., Gardner, E.L., 1993. Effects of haloperidol on human plasma magnesium. Journal of Psychiatric Research 27, 155–159. Kimmel, P.L., Peterson, R.A., Weihs, K.L., Simmens, S.J., Boyle, D.H., Verme, D., Umana, W.O., Veis, J.H., Alleyne, S., Cruz, I., 1995. Behavioral compliance with dialysis prescription in hemodialysis patients. Journal of the American Society of Nephrology 5, 1826–1834. Snowdon, J.A., Macfie, A.C., Pearce, J.B., 1976. Hypocalcaemic myopathy with paranoid psychosis. Journal of Neurology, Neurosurgery and Psychiatry 39, 48–52. Yassa, R., Nair, N.P., Schwartz, G., 1979. Plasma magnesium in chronic schizophrenia. A preliminary report. International Pharmacopsychiatry 14, 57–64.
Unexplained transient hyperphosphatemia is more common in acute psychiatric disorders than in acute medical–surgical conditions Ali Ahmadi, Edward C. Lauterbach*, Nasreen Malik, John Gonzales Department of Psychiatry and Behavioral Sciences, Mercer University School of Medicine, 655 First Street, Macon, Georgia 31201, USA Received 25 January 2001; received in revised form 18 October 2002; accepted 22 January 2003
Abstract Increased serum inorganic phosphorus associated with elevated serum calcium has been demonstrated to coincide with the onset of agitation and mania in periodic psychoses and bipolar disorders. We tested the hypothesis that unexplained transient hyperphosphatemia (UTHP) is more common in patients with psychiatric disorders than in controls with medical or surgical conditions. We studied 100 patients admitted to a psychiatric ward and 100 controls admitted to a medical–surgical ward. All subjects (patients and controls) underwent acute admission to the same general hospital. The serum phosphorus was measured upon admission and, if elevated, followed during the hospital course. Twenty patients (20%) with psychiatric disorders had unexplained hyperphosphatemia compared with four medical–surgical controls (4%). UTHP occurred in six patients with psychiatric disorders and no controls. Hypophosphatemia did not occur in subjects with psychiatric disorders. This study shows an increased incidence of UTHP in acutely ill, hospitalized patients with psychiatric disorders relative to acutely ill, hospitalized controls with medical–surgical conditions. These data extend previous findings by linking UTHP to acute psychiatric disturbances across varied psychiatric diagnoses independent of hypercalcemia. Potential explanations include trazodone administration and transient hypocalcemia. 䊚 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Phosphorus; Major depression; Schizophrenia; Bipolar disorder; Alcohol disorders; Substance abuse disorders
1. Introduction Elevated serum phosphorus levels accompanied by increased serum calcium have been observed *Corresponding author. Tel.: q1-473-301-2107; fax: q1478-301-5337. E-mail address: [email protected] (E.C. Lauterbach).
to coincide with exacerbations of rapid-cycling psychotic disorders and mania (Carman and Wyatt, 1979b). Additional studies have shown a significant reduction in agitation ratings in patients with agitated psychotic disorders following administration of probenecid (a potent phosphaturic agent) and calcitonin (Carman and Wyatt, 1979a,c). These increases in serum calcium and phosphorus, which precede behavioral changes, have been sug-
0165-1781/03/$ - see front matter 䊚 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0165-1781(03)00021-0
238
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
gested as a cause rather than an effect of psychotic agitation and mania (Snowdon et al., 1976). A significant number of patients with psychiatric disorders have disturbed behavior upon admission. We were struck by a clinical impression of remarkably frequent serum phosphorus elevations without calcium perturbations on our psychiatric ward and so we decided to study this phenomenon. Although previous studies evidenced a linkage between hyperphosphatemia and agitation associated with hypercalcemia observed in mania or psychosis (Snowdon et al., 1976; Carman and Wyatt, 1979a,b,c), medical–surgical controls were not employed. The aim of our study was to compare the serum phosphorus level of patients with psychiatric disorders to that of controls with medical or surgical conditions upon admission to the same hospital. We tested the hypothesis that unexplained transient hyperphosphatemia (UTHP) is more common in patients with psychiatric disorders than in control patients with medical or surgical conditions. 2. Methods We prospectively examined laboratory and clinical data on 100 consecutively admitted patients with psychiatric disorders and compared them to 100 consecutively admitted patients with medical– surgical conditions. All patients (psychiatric and medical–surgical) were admitted during the same time period, and patients with psychiatric disorders were admitted to a general psychiatric ward while patients with medical–surgical conditions were admitted to a medical–surgical ward. Both wards were part of a 519-bed community teaching hospital. Phosphorus levels for each group were determined on admission using conventional clinical laboratory techniques (Dryer and Routh, 1963) according to norms determined by the hospital clinical pathology department. Data collected for each subject included age, gender, psychiatric diagnoses, medicalysurgical diagnoses, all medications, their doses, routes and dosing intervals, phosphorus (normative reference level 2.4–4.6 mgydl), calcium, magnesium, blood urea nitrogen (BUN), creatinine, BUNycreatinine ratio, T3 uptake, T4, thyroid index, TSH and a
Table 1 Hyperphosphatemic etiologies Serum protein binding (lymphocyte dyscrasias) Decreased renal excretion Renal insufficiency Hypoparathyroidism Pseudohypoparathyroidism (Types I and II) Tumoral calcinosis Pseudoxanthoma elasticum Infantile hypophosphatasia Hyperostosis Hyperthyroidism Growth hormone Adrenal insufficiency Biphosphonate therapy Increased intestinal absorption Phosphorus-containing cathartics Medication with vitamin D compounds Granulomatous diseases producing vitamin D (sarcoidosis, tuberculosis) Internal redistribution Acute metabolic acidosis Acute respiratory acidosis Reduced insulin level Clonidine administration Cellular release Rhabdomyolysis Organ infarction Tumor lysis (lymphoma, metastatic small cell carcinoma) Thyrotoxicosis Acute hemolysis Parenteral administration Intravenous phosphate salts Phospholipid infusion Medical causes of hyperphosphatemia.
listing of all other abnormal laboratory and ancillary study findings. Parathormone levels were not obtained. Clinical charts were intensively reviewed for evidence of medical causes of phosphorus disturbances (Bennett and Plum, 1996). Hyperphosphatemia was considered to be unexplained if the chart review disclosed no evidence suggesting any of the possible etiologies of an elevated serum phosphorus level. A list of causes of hyperphosphatemia is provided in Table 1. We screened clinical data conservatively, construing even minor abnormalities on clinical exam or laboratory data to be suggestive of an explainable cause (e.g.
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
mildly elevated creatinine consistent with renal insufficiency). We assessed causation in all cases of disturbed phosphorus regardless of whether the abnormal value was transiently or persistently abnormal. Transient hyperphosphatemia was defined by the return of the phosphorus level back to the normal range on repeated determinations within 120 h of admission (evident on 2–5 repeated determinations), whereas persistent hyperphosphatemia was defined by consistently elevated values on repetitive determinations. Unexplained transient hyperphosphatemia (UTHP) was defined by absence of an explainable cause and transience, according to the above criteria. 2.1. Hypothesis testing Frequencies of hypophosphatemia, hyperphosphatemia, and UTHP were determined for the psychiatric and medical surgical groups. We hypothesized that UTHP is more common in patients with psychiatric disorders than in patients with medical–surgical conditions. We tested the hypothesis by comparing UTHP rates between the two groups using the one-tailed Pearson’s x2 with Yates correction factor. We assessed comparability of the two groups using a two-tailed two-way analysis of covariance of group and gender with age as a covariate. In the event of significance, a two-tailed two-way analysis of covariance of phosphorus by group with the significant findings as the covariate was determined. 2.2. Additional data exploration We subsequently explored the data for factors that might distinguish patients with UTHP from other patients by comparing patients with psychiatric disorders with UTHP to two additional sets of controls: (1) subjects with psychiatric disorders lacking UTHP; (2) subjects with psychiatric disorders lacking phosphorus aberration of any kind (i.e. lacking hyperphosphatemia, since hypophosphatemia was not observed in subjects with psychiatric disorders). We compared the subjects with UTHP to each of the two psychiatric control groups for demographic and clinical variables using the two-tailed Student’s t-test for analysis of
239
age and two-tailed Fisher’s exact test for categorical variables. 3. Results Initial inspection of laboratory values revealed hyperphosphatemic values in 22 psychiatric (ns 100) and 8 medical surgical (ns100) subjects. Means and standard deviations of phosphorus values for the groups were: UTHP (ns6 psychiatric subjects) 5.35"S.D. 0.3; No UTHP (ns94 psychiatric subjects) 4.01"0.81; Normal Phosphorus (ns76 psychiatric subjects) 3.84"0.50; Medical– Surgical Subjects (ns100) 3.43"0.66. Intensive chart review for causes of phosphorus disturbance revealed potential medical explanations in two psychiatric (hemolytic anemia, dialysis) and four medical–surgical (respiratory acidosis in two, diabetes mellitus, renal insufficiency) hyperphosphatemic cases. Hypophosphatemic values occurred in 0 psychiatric and 10 medical–surgical subjects, consistent with the somewhat lower mean phosphorus level in the medical–surgical group. Nine cases of hypophosphatemia in medical–surgical subjects revealed potential medical exlanations in each, including diuretic administration in four, malabsorption in two, alcoholism in one, insulin therapy in one, and depression (which was diagnosed as complicating the primary medical illness) in one. Thus, unexplained hyperphosphatemia (including UTHP cases) occurred in 20 psychiatric and four medical–surgical cases whereas unexplained hypophosphatemia occurred in no psychiatric and one medical–surgical subject (Table 2). Subjects with UTHP are displayed in Table 3. 3.1. Hypothesis testing UTHP was significantly more common in subjects with psychiatric disorders (Table 2) than in subjects with medical–surgical disease (6 of 100 psychiatric cases vs. 0 of 100 controls, one-tailed Yates-corrected Pearson’s x2 s4.2955, d.f.s1, P0.02). Subjects with psychiatric disorders (mean age 41.12"S.D. 13.65 years) were significantly younger than subjects with medical–surgical conditions (59.95"17.41, ANOVA Fs68.591, d.f.s 1, P-0.000000001). There was no significant
240
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
Table 2 Comparison of subjects with psychiatric disorders to medical–surgical controls Condition
Psychiatric subjects (ns100)
Medical–surgical controls (ns100)
Hyperphosphatemia Unexplained Transient (UTHP)* Hypophosphatemia Unexplained Transient Demographics Men** Women** Age†
22 20 6 0 0 0
8 4 0 10 1 0
50 50 40.64"S.D.13.49
42 58 60.03"S.D.16.92
*UTHP was significantly more common in subjects with psychiatric disorders than in controls (one-tailed Yates-corrected Pearson’s x2s4.2955, d.f.s1, P-0.02). **Gender difference was not significant between groups (Pearson’s x2 s1.284, d.f.s1, Ps 0.26). †Subjects with psychiatric disorders were significantly younger than controls (ANOVA Fs79.284, d.f.s1, Ps 0.0000000001).
difference in gender between the two groups (psychiatric: 50 males, 50 females; medical–surgical: 42 males, 58 females, Fs0.110, d.f.s1, Ps0.74). The analysis of covariance of phosphorus level by group with age as a covariate disclosed a significant effect of group (4.13"0.98 vs. 3.43"0.66, Fs17.137, Ps0.00005) but not of age (Fs2.713, d.f.s1, Ps0.1) on phosphorus level.
transient hypocalcemia (Ps0.09) compared to the ‘Normal Phosphorus’ group. In summary, UTHP was significantly more common in psychiatric patients than in medical-surgical patients and did not relate to other clinical variables, suggesting the possibility that UTHP may be linked to acute psychiatric conditions. 4. Discussion
3.2. Additional data exploration We explored the results for variables differing between subjects with UTHP and the two psychiatric control groups because larger samples in future studies might discriminate UTHP clinical features detectable only as non-significant trends in this study. Comparisons of the six subjects with UTHP to the two psychiatric control groups are displayed in Table 4. There were no significant differences (Bonferroni-corrected as0.0016) between subjects with UTHP and either control group for any of the demographic or clinical comparison variables (listed in Table 4). Without Bonferroni correction, there were non-significant trends toward younger age (Ps0.09) and more trazodone administration (Ps0.06) in the UTHP group compared to the ‘No UTHP’ group, and significantly more trazodone administration (Ps 0.04) and a non-significant trend toward more
This is the first report of a prospective controlled study demonstrating UTHP in acute, hospitalized subjects with psychiatric disorders when compared with acute, hospitalized subjects with medical– surgical disease. The significant difference in age between the two groups reflects the difference in age-related risks for psychiatric and medical–surgical conditions. Transient increases in serum phosphorus coinciding with elevated serum calcium and rapid cycling of psychotic agitation and mania were demonstrated in the past (Snowdon et al., 1976; Carman and Wyatt, 1979a,b,c). The aim of our study was to ascertain prospectively the serum phosphorus level of 100 psychiatric patients upon admission to the psychiatric ward, follow them up if levels were elevated during the hospital course, and compare them in the same fashion with 100 patients admitted to the medical–surgical ward. Previous studies did not use such controls and
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
241
Table 3 Characteristics of patients with psychiatric disorders and UTHP Case
Age
Sex
Diagnosis
Medications
1
41
F
Major depression
2
38
M
Cocaine-induced psychotic disorder
3
22
M
Substance-induced psychotic disorder
4
42
M
Schizophrenia chronic paranoid type
5
31
F
6
32
M
Alcohol-induced mood and psychotic disorders Adjustment disorder with depressed mood Schizoaffective disorder
Bupropion 100 mg Nitroglycerine Risperidone 6 mg Fluoxetine 60 mg Trazodone 100 mg Risperidone 4 mg Valproic acid 1000 mg Trazodone 100 mg Fluphenazine decanoate 50 mgymo Benztropine 2 mg Trifluoperazine 100 mg Ibuprofen 600 mg
Quetiapine 800 mg Venlafaxine 75 mg Diphenhydramine 100 mg Trazodone 200 mg Benztropine 2 mg Haloperidol 15 mg Valproic acid 1000 mg
The age, gender, diagnosis at discharge, and medication regimen are displayed for each subject with UTHP.
Table 4 Comparison of subjects with UTHP to other subjects with psychiatric disorders
Male Female Age Psychotic diagnosis Psychotic condition Mood diagnosis Mood condition Antidepressants Antipsychotic Valproic acid Benztropine Trazodone Lithium Substance abuse Transient hypocalcemia Elevated T3 uptake test
UTHP (ns6)
No UTHP (ns94)
Normal phosphorus (ns78)
4 2 34.33qS.D.7.55 1 4 2 3 4 4 2 2 3 0 3 3 3
46 48 40.95qS.D.13.67† 30 38 44 48 35 41 16 9 15† 6 28 15 24
36 42 42.11qS.D.13.88 24 31 36 38 28 33 14 7 10* 5 23 15† 21
Comparison of 6 subjects with psychiatric disorders and UTHP to the 94 subjects with psychiatric disorders lacking UTHP (‘No UTHP’) and the 78 subjects with psychiatric disorders lacking phosphorus aberrations of any kind (‘Normal Phosphorus’). †P0.1, *P-0.05.
242
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
their main objective was to demonstrate transient hypercalcemia as a trigger of the agitated state observed in rapidly cycling acute exacerbations of mania or psychosis. In contrast with previously reported studies, we did not detect transient hypercalcemia in our patients with psychiatric disorders (Snowdon et al., 1976; Carman and Wyatt, 1979a,b,c). Rather, we identified hyperphosphatemia independent of hypercalcemia. Although our study relied on the prospective identification of UTHP in 100 patients with psychiatric disorders and 100 patients with medical– surgical illness, and used conventional standardized laboratory assessment, certain methodological limitations merit consideration. These limitations include the use of clinical diagnoses (as opposed to structured clinical diagnostic interviews), the possibility of undetected physical illness, age and medication differences between the groups, possible laboratory error, and insufficient power to detect differences between UTHP and non-UTHP psychiatric subgroups. The use of clinical rather than structured interview diagnoses, however, does not alter our UTHP finding, ensures comparability with controls (medical–surgical patients were also diagnosed clinically), and allows generalization of our findings to other clinical populations. Undetected physical illness in subjects with psychiatric disorders remains a possibility although this is less likely in light of the transiency of UTHP, younger age of these patients with psychiatric disorders, and reciprocal cross-training of each physician on both wards used in the study. We chose medical–surgical subjects from the same hospital as a comparison group because this group would be expected to have high rates of phosphorus abnormalities and constitute a rigorous test of the hypothesis, but we did not anticipate such a significant difference in subject’s age. Nevertheless, neither younger age nor the medications in our subjects with psychiatric disorders are known to increase adult phosphorus levels, and phosphorus levels were not significantly related to age in our subjects. Clinical laboratory false-positive rates for hyperphosphatemia would be expected to be identical for the two groups since levels were determined over the same period of time using the same assay. Moreover, false-positive hyperphos-
phatemia rates would be expected to be higher in the medical–surgical patients since these patients often undergo a greater number of laboratory tests, thus multiplying the chances of laboratory error and thereby biasing this study against finding an increase in subjects with psychiatric disorders. Finally, we did not originally design this study with the intent of looking for predictive differences between UTHP and non-UTHP psychiatric subgroups. Overall, the use of our medical–surgical control group with its possible bias toward increased UTHP makes it likely that the increased rate of UTHP in subjects with psychiatric disorders is a legitimate finding. There has been some previous speculation in the literature regarding ionic involvement in the pathophysiology of psychiatric illnesses. In one study, no difference in serum phosphorus level was found between patients with chronic schizophrenia and patients with non-psychotic mental retardation (Yassa et al., 1979). In another study, haloperidol treatment did not reduce the serum levels of phosphorus or calcium but reduced the magnesium level (Jabotinsky-Rubin et al., 1993). Interestingly, among 88 patients with end-stage renal disease, higher phosphorus levels were present in patients undergoing hemodialysis who were younger and more depressed in comparison to other hemodialysis patients (Kimmel et al., 1995). In the same study higher phosphorus levels correlated with aberrant behavior and poor dialysis compliance (Kimmel et al., 1995). Metabolic stress associated with acute psychiatric exacerbation or membrane phospholipid metabolic disturbances (Fukuzako et al., 1999) in our patients might possibly relate to transient hyperphosphatemia. A transient endocrinologic or metabolic disturbance is suggested by the over-representation of transient hypocalcemia in psychiatric subjects with UTHP relative to psychiatric subjects with normal phosphorus. It is also possible that trazodone administration may predispose to UTHP since trazodone administration was over-represented in UTHP subjects, although a MEDLINE literature search did not disclose an association. Alternatively, elevated phosphorus may reflect otherwise subclinical rhabdomyolysis, similar to
A. Ahmadi et al. / Psychiatry Research 117 (2003) 237–243
increased serum creatine phosphokinase seen in psychotic exacerbations. Although there was no evidence of clinical rhabdomyolysis in UTHP subjects, one had an elevated SGOT and another had an elevated LDH. In any event, these data extend previous findings by linking UTHP to acute psychiatric disturbances across various psychiatric diagnoses independent of persistent hypercalcemia. It will now be of interest to study patients compared to controls matched for age, gender, and extent of medical illness. References Bennett, J.C., Plum, F., 1996. Cecil Text Book of Medicine. 21st ed. WB Saunders Co, Philadelphia. Carman, J.S., Wyatt, R.J., 1979a. Calcium: bivalent cation in the bivalent psychoses. Biological Psychiatry 14, 295–336. Carman, J.S., Wyatt, R.J., 1979b. Calcium: pacesetting the periodic psychoses. American Journal of Psychiatry 136, 1035–1039.
243
Carman, J.S., Wyatt, R.J., 1979c. Use of calcitonin in psychotic agitation or mania. Archives of General Psychiatry 36, 72– 75. Dryer, R.L., Routh, J.I., 1963. Determination of serum inorganic phosphorus. Clinical Chemistry 4, 191–197. Fukuzako, H., Fukuzako, T., Kodama, S., Hashiguchi, T., Takigawa, M., Fujimoto, T., 1999. Haloperidol improves membrane phospholipid abnormalities in temporal lobes of schizophrenic patients. Neuropsychopharmacology 21, 542–549. Jabotinsky-Rubin, K., Durst, R., Levitan, L.A., Moscovich, D.G., Silver, H., Lerner, J., Van Praag, H., Gardner, E.L., 1993. Effects of haloperidol on human plasma magnesium. Journal of Psychiatric Research 27, 155–159. Kimmel, P.L., Peterson, R.A., Weihs, K.L., Simmens, S.J., Boyle, D.H., Verme, D., Umana, W.O., Veis, J.H., Alleyne, S., Cruz, I., 1995. Behavioral compliance with dialysis prescription in hemodialysis patients. Journal of the American Society of Nephrology 5, 1826–1834. Snowdon, J.A., Macfie, A.C., Pearce, J.B., 1976. Hypocalcaemic myopathy with paranoid psychosis. Journal of Neurology, Neurosurgery and Psychiatry 39, 48–52. Yassa, R., Nair, N.P., Schwartz, G., 1979. Plasma magnesium in chronic schizophrenia. A preliminary report. International Pharmacopsychiatry 14, 57–64.