Theophylline

Limits on what this shows

• Null and negative findings: Pyridoxal and 4-pyridoxic acid unchanged [5]. In the 23-child cohort, vitamins A, B2, B12 and C were not affected [10]. Nine weeks of oral pyridoxine in steroid-dependent asthma produced no benefit on asthma outcomes (31 patients), so B6 repletion is not shown to improve asthma control [13]. Pyridoxine for theophylline-induced seizures lacks significant supporting data [17].
• At-risk subgroups: Children on long-term therapy or at 5 to 15 mcg/mL [11]; older adults with COPD, who are already often subnormal without the drug (56 to 70%) [12].

Direction

Depletion of the circulating active cofactor (PLP). Plasma pyridoxal and urinary 4-pyridoxic acid were unchanged in the Pretoria studies, so the signal is specific to the PLP pool rather than a fall in total B6 intake [5][6].

Regulatory status

Not mentioned in the label sections read [1]. The NIH Office of Dietary Supplements states that patients treated with theophylline often have low plasma PLP, which could contribute to neurological side effects [18]; the Linus Pauling Institute states that PLP bioavailability may be reduced by methylxanthines such as theophylline [19]. Both are Tier B references, not primary data.

Source 1: Pretoria, drug isolated

Placebo-controlled, double-blind study in apparently healthy young men [6]: theophylline significantly lowered plasma PLP with plasma pyridoxal unchanged; erythrocyte PLP also fell (p = 0.03); urinary xanthurenic acid after a tryptophan load rose (p = 0.007) after 4 weeks; one week of pyridoxine at 10 mg/day (experimental dose in healthy men; NIH adult upper level is 100 mg/day [18]) normalised plasma and erythrocyte PLP and the tryptophan load result. The same group reported lower PLP in 28 asthmatic women than 33 controls (p < 0.001) and large PLP falls in 17 volunteers given theophylline [5], and PLP falls with a threefold rise in erythrocyte pyridoxal kinase in seven healthy men [7]. Counted here as one source lineage because authors and volunteer pools overlap.

Source 2: Shimizu, children, Japan

26 asthmatic children, 20 on slow-release theophylline and 6 on none [9]: PLP 5.3 vs 9.0 ng/mL (p < 0.05); PLP inversely related to serum theophylline (rs = -0.609, p < 0.001); a single 200 mg oral dose lowered PLP in five children at 4 hours (p < 0.05) while pyridoxal did not change. A second cohort of 23 hospitalised children (16 on theophylline, 7 not) [10]: vitamin B6 7.1 vs 11.8 ng/mL (p < 0.05), rs = -0.657 (p < 0.001). The authors did not report a CI; if the plus-minus values are SD, a Welch 95% CI for the difference of -4.7 ng/mL is -6.6 to -2.8 (derived here; if they are SEM the interval would cross zero and contradict the reported p, so the dispersion type is unverified).

Source 3: Tanaka, children, Japan

26 asthmatic children on sustained-release theophylline [11]: PLP correlated negatively with serum theophylline (r = -0.59); PLP was lower after more than 5 weeks than after less than 4 weeks (p < 0.05) and lower at 5 to 15 mcg/mL than below 5 mcg/mL (p < 0.05). Different authors from Source 2; patient overlap is not stated.

Source 4: Martinez de Haas, Netherlands

Whole-blood PLP below the reference range (35 to 107 umol/L): about 94% of 17 geriatric COPD patients on theophylline and 96% of 25 younger COPD patients on theophylline, vs 70% of 40 geriatric COPD patients not on it and 56% of 84 geriatric non-COPD patients [12]. Derived here from counts back-calculated from the rounded percentages: pooled theophylline users 40 of 42 (95%) vs untreated COPD 28 of 40 (70%), risk difference 25 points (Newcombe 95% CI 9 to 41), odds ratio 8.6 (Woolf 95% CI 1.8 to 41.3). The geriatric-only contrast (16 of 17 vs 28 of 40) is weaker (odds ratio 6.9, 95% CI 0.8 to 57.7). The authors report no CI.

Source 5: Sur, adults, United States

In 31 steroid-requiring adults with asthma, baseline plasma pyridoxine was lower in theophylline users than non-users, 10 +/- 8 vs 19 +/- 6 ng/mL (p =.01); group sizes were not given [13].

Functional relevance

In a double-blind trial (20 enrolled, 15 completers) B6 supplementation reduced theophylline-related tremor after a single dose (p < 0.01); most other measures did not differ from placebo [14].

Mechanism

Present, with unresolved detail. Theophylline is a noncompetitive inhibitor of erythrocyte pyridoxal kinase (apparent Ki 1.28 x 10^-5 mol/L, about 13 umol/L, below the roughly 55 to 110 umol/L implied by therapeutic serum levels of 10 to 20 mg/L) and does not affect pyridoxamine-5'-phosphate oxidase [6][1]. In vivo, erythrocyte kinase activity roughly doubles (24.2 +/- 4.0 to 46.9 +/- 7.3 nmol PLP/g Hb/h) [6] and tracks drug level independently of B6 status (r = 0.71, p < 0.001) [8], so the net result is a lower PLP despite a compensatory enzyme rise. Rabbit data show a 47% fall in plasma PLP and suggest ethylenediamine in aminophylline can bind PLP [15]; colonocyte pyridoxine uptake was not inhibited by theophylline [16].

Counter-evidence held on record

Counter-evidence and confounding Confounding by indication is real: asthma itself is associated with lower PLP (28 asthmatic women vs 33 controls, p < 0.001 [5]; 56 to 70% subnormal without theophylline [12]), and the patient datasets are small, unadjusted and not controlled. Causal attribution to the drug rests on the placebo-controlled healthy-volunteer work [6] plus within-patient dose-response [9][10][11]. The Pretoria papers share authors and volunteers, so the independent lineages are five (Pretoria, two Japanese pediatric groups, Dutch, American), not ten papers. Effect sizes are mostly p-values; no author-reported CIs. Full texts were not opened. Default governance for red applies ( ); under the governance table it advances to only after an editor verifies the second independent Tier A source and signs off. Framing caution This is a circulating cofactor marker, not proof of dietary B6 inadequacy, and it does not show that supplementation changes asthma or seizure outcomes. No supplement dose is recommended here. Any B6 supplement decision belongs to a prescriber or pharmacist.

Limits on what this shows

• Null and negative findings: Chronic overmedication (n = 3) did not lower K [20]. No additive effect with inhaled beta-2 agonists in one study [22] but potentiation in another [23], so the interaction is inconsistent while the direction of the theophylline effect is not. No study showing an unchanged potassium with theophylline alone was found.
• At-risk subgroups: Serum levels above 30 mcg/mL and acute overdose; rapid IV loading; co-treatment with beta-2 agonists; children on IV courses.

Direction

Disturbance by transcellular shift (serum level falls). Not a dietary deficiency and not evidence that intake should rise (Hard Rules 2 and 5).

Regulatory status

The label's Overdosage Table IV reports hypokalemia in 85% (acute study 1, n = 157), 79% (acute study 2, n = 14), 44% (chronic study 1, n = 92) and 43% (chronic study 2, n = 102) of patients with serum theophylline above 30 mcg/mL, and directs that serum electrolytes and glucose be measured on presentation [1].

Population evidence: toxic range

Shannon: acute intoxication (n = 10, peak theophylline 86.6 mcg/mL) had serum K 2.80 +/- 0.16 mEq/L vs 4.10 +/- 0.70 in chronic overmedication (n = 3, peak 73.0 mcg/mL; p <.0001), the chronic group being similar to healthy controls (n = 6); epinephrine was higher in the acute group (282 vs 133 vs 58 pg/mL) [20].

Population evidence: broad range

Flack: 869 patients with measurable theophylline vs 350 controls without airway disease, hypokalemia unadjusted OR 4.2 (95% CI 2.2 to 7.9), with a linear concentration-response from about 5.5 umol/L (1 mg/L) upward [25].

Population evidence: therapeutic range

In normal volunteers, IV theophylline alone at a mean maximal 14.9 +/- 2.2 mg/L produced a small but significant fall in plasma potassium, with no additive effect on inhaled fenoterol hypokalemia [22]; seven COPD patients given IV theophylline 6 to 7 mg/kg had a significant potassium fall within the 10 to 20 mg/L range (letter excerpt only) [21]; serum K fell in 46 asthmatic children on an IV course (p < 0.01 vs controls; serum theophylline not reported) [24]. In normal subjects at therapeutic plasma theophylline, the serum potassium fall during an IV terbutaline infusion was significantly greater with theophylline present [23]. IV theophylline also blunted the normal morning-to-noon serum K rise in 10 women [26].

Mechanism

Named. Beta-2 adrenergic stimulation by raised catecholamines: the potassium and glucose differences between acute and chronic overdose paralleled epinephrine levels [20]. The label does not state a mechanism [1]. Background pharmacology is nonselective PDE inhibition and adenosine antagonism [4].

Counter-evidence held on record

Counter-evidence and confounding Hard Rule 2 applies: this is the drug's toxicity and pharmacologic effect, labeled as a disturbance. In Flack the ORs are unadjusted against controls with no airway disease, so beta-agonist, steroid and diuretic co-treatment are uncontrolled confounders at therapeutic levels; the overdose cohorts, Shannon's controlled comparison and the acute IV studies carry the causal weight. The therapeutic-range effect is small and several records are listings or letters only. Hypokalemia can reverse as the shift resolves. Framing caution Excess/redistribution framing: serum potassium falls because potassium moves into cells. Potassium supplementation is a decision for clinicians managing high theophylline levels, never a consumer remedy from this entry.

Limits on what this shows

• Null and negative findings: None found in stable therapy cohorts, but the lower CI bound of Flack's OR is 1.0, and no controlled trial of serum Mg at therapeutic levels was located.
• At-risk subgroups: Supratherapeutic levels, IV courses, co-treatment with beta-2 agonists and corticosteroids, older adults on long-term therapy.

Direction

Depletion of serum magnesium, partly renal (magnesuria) and partly shift.

Regulatory status

The label's overdose section lists decreases in serum phosphate and magnesium among toxicity manifestations; no figures are given [1].

Population evidence

Flack: hypomagnesemia unadjusted OR 1.6 (95% CI 1.0 to 2.5) in 869 patients vs 350 controls, with a linear concentration-response [25]. Das: 14 of 50 stable-asthma outpatients (28%) were hypomagnesemic, a status significantly associated with theophylline use among several other factors (45 healthy controls; no ORs in the record) [27]. Overdose series: 64 poisonings, hypomagnesaemia among common abnormalities, count not given [28]; 22 overdoses, common in the subset measured [29]. One case of serum Mg 1.1 mEq/L after 20 years of therapy, corrected by oral magnesium, link suspected not proven [31].

Mechanism

Present. IV theophylline raised 5-hour urinary Mg from 0.54 +/- 0.07 to 0.86 +/- 0.10 mmol (n = 10; the authors infer intracellular depletion) [26]. In ewes theophylline caused hypomagnesaemia with magnesuria that the authors judged mainly extrarenal [30].

Counter-evidence held on record

Counter-evidence and confounding The therapeutic-range signal is confounded by indication and co-medication: Flack's controls had no airway disease and the ORs are unadjusted, and Das is a cross-sectional comparison with many co-medications. Das also linked low magnesium to long-acting beta-2 agonist and inhaled corticosteroid use, so co-medication is a competing explanation [27]. Orange rather than red because the effect at therapeutic levels is modest, the CI touches 1.0, and no controlled therapeutic-range measurement exists. Boundary to display with the claim: supratherapeutic or IV exposure; association in stable asthma. Framing caution A serum disturbance associated with drug level, not a statement that dietary magnesium is low. Do not recommend magnesium supplementation from this entry.

Limits on what this shows

• Null and negative findings: None found; no study of phosphate at stable therapeutic levels with adjustment was located.
• At-risk subgroups: Children and adults receiving IV courses; high serum levels.

Direction

Disturbance (serum fall); mechanism inferred as a shift, not shown for phosphate.

Regulatory status

Overdose section lists decreases in serum phosphate [1].

Population evidence

Flack: hypophosphatemia unadjusted OR 2.7 (95% CI 1.2 to 5.3), concentration-related [25]. Amin: serum phosphate fell after therapy vs controls (p < 0.01) in 21 children with acute and 25 with chronic asthma on IV theophylline; serum theophylline not reported [24]. Knutsen: IV theophylline accentuated the normal morning-to-noon fall in serum phosphate in 10 women (values -0.13 +/- 0.04 to -0.23 +/- 0.03 mmol/L, but the record's pairing of values may be misassigned) [26].

Mechanism

Absent as a source-stated mechanism for phosphate. The beta-2 adrenergic explanation documented for potassium [20] is a plausible parallel but was not tested for phosphate.

Counter-evidence held on record

Counter-evidence and confounding Same confounding as the other electrolyte entries (unadjusted controls without airway disease; co-treatment with beta-2 agonists). Two of the three sources lack drug concentrations or have uncertain value pairings. Orange rather than red because magnitude at therapeutic levels is unclear and the mechanism for phosphate is not source-stated. Boundary to display: IV or supratherapeutic exposure. Framing caution A serum disturbance in monitored patients. This entry does not support phosphate supplementation.

Limits on what this shows

• Null and negative findings: The direction is not consistent across Tier A cohorts: serum calcium rose in toxicity [32][1] but fell in the pediatric acute-asthma cohort [24] and in three toxic patients [32]. Flack did not report calcium [25].
• At-risk subgroups: Patients with toxicity, older adults (single case), and people receiving IV courses.

Direction

Excess (serum rise) in toxicity; urinary calcium loss in one IV study; a serum fall in one pediatric cohort. Mixed and labeled as such.

Regulatory status

The label's overdose section lists increases in serum calcium among toxicity manifestations [1].

Population evidence

McPherson: of 60 hospitalised patients with theophylline toxicity over two years, 11 were hypercalcemic, resolving as levels fell; calcium fell significantly in 3 others; a significant calcium rise at therapeutic levels in normal volunteers was reversed by propranolol (no values in the record) [32]. Case: 90-year-old woman, theophylline 16.6 mcg/mL, hypercalcemia with hypokalemia and acidosis [33]. Knutsen: 5-hour urinary calcium 0.89 +/- 0.18 to 1.45 +/- 0.26 mmol on IV theophylline (n = 10), read by the authors as a negative calcium balance [26]. Amin: total serum calcium decreased in acute asthma vs controls (p < 0.01) [24].

Mechanism

Named for the serum rise: a beta-adrenergically regulated system, because propranolol reversed it [32].

Counter-evidence held on record

Counter-evidence and confounding The pediatric fall and the urinary loss run opposite to the toxic-range serum rise, acute illness itself alters calcium, and McPherson's record carries no values. Rat bone and vitamin D data are kept in a separate yellow entry. Orange because one well-bounded Tier A cohort shows a real, reversible serum effect; not red because direction is mixed and numbers are missing. Boundary to display: toxicity and short-term exposure. Framing caution Excess or retention-type framing: the hazard is raised serum calcium, so calcium or vitamin D supplements are not a remedy suggested by this entry.

Limits on what this shows

• Null and negative findings: None found; no second cohort with serum sodium located.
• At-risk subgroups: Hospitalised asthma and COPD patients (the setting of the cohort); co-treatment was not reported in the record.

Direction

Disturbance (serum fall) and urinary sodium loss; not a dietary matter.

Regulatory status

Not mentioned in the label sections read [1].

Population evidence

Flack: hyponatremia unadjusted OR 5.4 (95% CI 2.0 to 12.9), concentration-related, 869 vs 350 controls [25]. Knutsen: urinary sodium 22.9 +/- 7.5 to 49.4 +/- 9.5 mmol over 5 hours on IV theophylline (n = 10) [26]. Amin measured sodium but the record reports no result [24].

Mechanism

Partly present as natriuresis [26]; not shown to explain a fall in serum sodium.

Counter-evidence held on record

Counter-evidence and confounding Rule 6: confounding by indication is presumed. Controls had no airway disease, ORs were unadjusted, and diuretics, illness and fluids are plausible causes. Yellow because causal attribution is not secured, although the Tier A study exists; it could rise to orange with an adjusted analysis or a second cohort. Framing caution Monitoring prompt in hospitalised patients, not a deficiency claim. Do not recommend salt or fluid changes from this entry.

Limits on what this shows

• Null and negative findings: No B1 relationship to theophylline concentration [10]. Vitamins A, B2, B12 and C were unaffected in the same cohort [10]. No second cohort located.
• At-risk subgroups: Not defined.

Direction

Depletion direction, single small cohort.

Regulatory status

Not mentioned in the label sections read [1].

Population evidence

Shimizu: B1 38.4 +/- 1.6 vs 46.4 +/- 3.5 ng/mL (p < 0.05; 16 on slow-release theophylline vs 7 on none), with no relationship between theophylline and B1 levels; the authors called a dose-dependent interaction unlikely [10].

Mechanism

Absent. No pathway linking theophylline to thiamine handling was found.

Counter-evidence held on record

Counter-evidence and confounding The same blood samples gave the B6 result, so this entry shares a cohort with the B6 entry (see Overlap map). Hospitalised, non-randomised comparison with seven controls; the vitamin difference could reflect illness or diet. The absence of a dose-response and of a mechanism keeps causal attribution too weak for orange. Framing caution Not a deficiency claim. Do not recommend thiamine supplementation from this entry.

Limits on what this shows

• Null and negative findings: None found.
• At-risk subgroups: People with diabetes; acute overdose; hospitalised patients on IV courses.

Direction

Excess (glucose raised). A pharmacologic or hormonal effect, never a nutrient deficiency (Hard Rule 2).

Regulatory status

Label: theophylline at 10 to 20 mcg/mL modestly increases plasma glucose (mean 88 to 98 mg%); uric acid, free fatty acids, total cholesterol and urinary free cortisol also rise. Overdose Table IV: hyperglycemia in 98% of acute study 1 and 18% of chronic study 1 [1].

Population evidence

Shannon: glucose 208 +/- 26 mg/dL in acute intoxication vs 139 mg/dL in chronic overmedication (p <.04), paralleling epinephrine [20]. Flack: hyperglycemia unadjusted OR 2.3 (95% CI 1.7 to 3.0) [25].

Mechanism

Named: catecholamine-linked, as for potassium [20].

Counter-evidence held on record

Counter-evidence and confounding Strong Tier A support, but blood glucose is not a nutrient status and the effect is the drug's own pharmacology. It is graded yellow as a context flag rather than orange or red to avoid filing a pharmacologic effect as a deficiency-type finding. Framing caution Monitoring item for people with diabetes. No nutrient or supplement implication.

Limits on what this shows

• Null and negative findings: A PubMed search for theophylline with bone mineral density, osteoporosis or 25-hydroxyvitamin D (48 records) returned no human cohort measuring these after theophylline in the first 10 records; the label does not mention bone.
• At-risk subgroups: Not defined.

Direction

Depletion direction, hypothesized from an animal model.

Regulatory status

Not mentioned in the label sections read [1].

Population evidence

Rat study (young males, continuous subcutaneous infusion, 4 weeks): urinary calcium higher (p < 0.0001), total body calcium per gram body weight lower (p < 0.05), serum 25-hydroxyvitamin D lower (p < 0.002) [34]. Human, indirect: urinary calcium rose from 0.89 +/- 0.18 to 1.45 +/- 0.26 mmol over 5 hours on IV theophylline in 10 women, read by the authors as negative calcium balance [26].

Mechanism

Proposed, not shown: hepatic microsomal enzyme induction speeding vitamin D metabolism, as with barbiturates and phenytoin [34]. The extracted label text does not describe theophylline as an enzyme inducer [1].

Counter-evidence held on record

Counter-evidence and confounding Animal dosing and serum levels may not reflect human exposure, and the human urinary-calcium data are acute. Corticosteroids, immobility and illness are competing explanations for low bone density in people with asthma. Shares the urinary calcium finding with the calcium entry. Framing caution Not a deficiency claim. No vitamin D or calcium supplementation is suggested.