RESULTS
From over 2,000,000 cases in the database, 3541 lithium cases originated from health care facilities, 347 had hemodialysis recommended or performed, and 298 recorded both a supratherapeutic lithium concentration and new neurological symptoms. Following exclusion for incomplete data, 219 patients were analyzed.(Table 2) The Paris criteria could be applied to 70 patients, while one or both EXTRIP criteria could be applied to 178 patients (78 “recommended”; 146 “suggested”). Forty-two patients were excluded from the comparison because data were insufficient to determine whether Paris criteria could be applied.
Seventy patients met PARIS criteria for hemodialysis. (Table 3) Three deteriorations (all of whom died) occurred among the 65 patients who had hemodialysis. No deteriorations occurred in the 5 patients who did not have hemodialysis. Statistical analysis was not performed because no events occurred in one cell. One-hundred and seven patients did not meet Paris criteria for hemodialysis. Deteriorations occurred in 2/83 patients (2.4%) dialyzed as opposed to 4/24 patients (16.6%) not dialyzed (p=0.022) The odds ratio of deterioration for patients who did not meet PARIS criteria for hemodialysis and were not dialyzed was 8.1 (95% CI 1.4-47.4).
Table 3 shows the data, analysis and comparisons for patients who met the EXTRIP suggests criteria. As noted above, many statistical comparisons were not possible because of zero events occurring in at least one cell. Only two analyses reached statistical significance. Among the 107 patients who did not meet the Paris criteria for hemodialysis, 83 underwent the procedure. Neurological deterioration was more likely in the remaining 24 patients who were not dialyzed (p = 0.22; OR 8.10(95% CI 1.39 - 47.39). For 81 patients outcomes could be compared when EXTRIP criteria were met and PARIS criteria were not. Clinical deterioration was more likely in these patients when they did not undergo hemodialysis (p=0.01, OR 8.67 95% CI: 1.42 - 52.94). When both criteria were met the majority of patients were dialyzed and there were no deteriorations. Similarly, when both criteria were not met, the majority of patients were dialyzed and there were no deteriorations.
Table 4 shows the data and analysis and comparisons for patients who met the EXTRIP recommend criteria. As noted above, many comparisons were not possible because of zero events occurring in at least one cell. The comparison between EXTRIP recommends and PARIS did not reach statistical significance.
There were 181 patients who met either the recommended or suggested EXTRIP criteria for hemodialysis. Their analysis is summarized in table 5. Four of the 26 (15.4%) who did not have hemodialysis deteriorated and one patient died as opposed to the 151 patients who underwent hemodialysis among whom 5 deteriorated (3.3%) and 3 died (p=0.031). The odd ratio for deterioration in the patients not dialyzed was 5.1 (95% CI 1.3-20.3). In 41 patients EXTRIP neither recommended nor suggested hemodialysis. All patients did well including 5/41 who did not receive hemodialysis. Statistical analysis was not performed because no events occurred in one cell. For 62 patients both the PARIS criteria and at least one of the EXTRIP criteria for hemodialysis were met. The only three deteriorations were among the 57 patients who underwent hemodialysis. Statistical analysis was not performed because no events occurred in one cell. There were 21 patients who fulfilled neither criteria for hemodialysis. There were no deteriorations in this group, including the five patients who did not receive hemodialysis. Statistical analysis was not performed because no events occurred in one cell. There were 86 patients who met at least one of the two EXTRIP criteria for hemodialysis but did not meet the Paris criteria. Hemodialysis was performed in 67 patients two of whom (3.0%) deteriorated. In contrast, among patients not dialyzed 4/19 (21.1%) deteriorated, one of whom died (p=0.019). The odd ratio for deterioration if dialysis was not performed when EXTRIP was in favor of hemodialysis and PARIS was not was 8.7 (95% CI 1.5-51.8). Table 6 provides a more detailed description of these four patients. Finally, there were 8 patients for whom PARIS was in favor of hemodialysis but neither of the two EXTRIP criteria were met. All 8 patients underwent hemodialysis and there were no deteriorations. Statistical analysis was not performed because no events occurred in one cell.