Increased Incidence of Pseudotumor Cerebri Syndrome Among Users of Tetracycline Antibiotics

INTRODUCTION

Pseudotumor cerebri syndrome (PTCS) is characterized by increased intracranial pressure with normal brain parenchyma, absence of hydrocephalus, mass lesion, or underlying infection or malignancy.¹ PTCS can be classified as primary, i.e. idiopathic intracranial hypertension (IIH), or secondary.^{2, 3}

Tetracyclines are a class of antibiotics that have long been considered to be a precipitating factor for pseudotumor cerebri syndrome.^4–8 Patients typically present with headache, transient visual obscurations, diplopia, and pulsatile tinnitus.¹ While a causative link between tetracycline use and PTCS has yet to be firmly established, studies have recognized the onset of PTCS in patients who use tetracycline antibiotics, as well as improvement of symptoms following discontinuation of the medication.^7–11 The course of tetracycline-induced pseudotumor cerebri syndrome (PTC-T), however, is variable in terms of duration of antibiotic use prior to the onset of symptoms and duration of symptoms following discontinuation of the antibiotic.^{9, 12} We recently compared the clinical features of PTC-T and IIH and found that while there were many similarities, there were also important differences.¹³ Specifically, we found that PTC-T patients were younger and were less likely to be obese than those diagnosed with IIH. We also found that the duration of illness tended to be shorter and recurrence rates tended to be lower in the PTC-T group compared to the IIH group.

The incidence of PTC-T is not known. A comparison of the incidence of PTCS among tetracycline-users and the incidence of idiopathic intracranial hypertension (IIH) among the general population could potentially strengthen the suggestion of causation.^{14, 15} This information could also aid providers in the decision to prescribe tetracycline antibiotics and help guide patient education regarding adverse drug reactions.

METHODS

We searched the University of Utah Health system’s records to estimate the incidence of PTC-T. We created a REDCap (Research Electronic Data Capture) database and populated it by searching records for a diagnosis of pseudotumor cerebri or idiopathic intracranial hypertension (ICD code 348.2 for dates of service before 2015). This query provided us with a list of all patients who were diagnosed with all-cause PTCS and IIH between 2007 and 2014. We confirmed that each patient had been evaluated by and diagnosed with PTCS by a University of Utah neuro-ophthalmologist. We then searched this group for patients who had been diagnosed with PTC-T. We specifically limited our search to those patients between 12 and 50 years of age.

We used the Federated Utah Research and Translational Health e-Repository (FURTHeR) query tool ¹⁶ to query the University of Utah Enterprise Data Warehouse for patients between the ages of 12 and 50 who were prescribed tetracycline-class antibiotics between 2007 and 2014. This database aggregates all clinical data from electronic medical records within the University of Utah health system. We then estimated the incidence of PTC-T per 100,000 person-years by dividing the number of patients with PTC-T by the number of patients who had been prescribed a tetracycline antibiotic, dividing by the eight years covered by the study, and then multiplying by 100,000.

Statistical and Sensitivity Analysis

To compute the incidence of tetracycline-induced pseudotumor cerebri syndrome, we required an accurate numerator (the number of cases) and an accurate demoninator (the number of patients at risk) for the State of Utah. The University of Utah is the only tertiary ophthalmology center in Utah, so we assumed our numerator, the number of cases in the University of Utah Health system, would be equal to the total number of cases in the State of Utah. We could estimate the number of patients at risk in the University system as the total number of patients with a University-system tetracycline prescription. We obtained the Utah population from the US Census. By assuming that the patients seen in the University system are representative of the patients in the entire state of Utah, and knowing the portion of the Utah population who receives their healthcare in the University system, we could estimate the number of patients in the entire state who had a tetracycline prescription. This number would then provide a statewise denominator of patients at risk. Because of the small number of cases, we multiplied the number of cases by 100,000 to obtain the one-year incidence per 100,000 patients, which is the same as cases per 100,000 person-years. Because our calculation required a statewide denominator that was not directly derived from individual patient prescription data, we termed this portion of the study a sensitivity analysis, where the accuracy of the calculation depends on the accuracy of the assumptions.

RESULTS

For the period 2007–2014, our database contained a total of 274 patients diagnosed with PTCS or IIH. Of these 274 patients, 45 patients had a new diagnosis of PTC-T (16.4%). Forty of these 45 patients listed a Utah zip code in their home address. Six patients had been prescribed doxycycline, 37 had been prescribed minocycline and 2 had been prescribed tetracycline.

At the end of 2014, the FURTHeR query contained records for 2,547,636 patients. For the time period 2007–2014, 876,358 of these patients were between the ages of 12 and 50 years. Of these, 960 (0.11%) received a new prescription for a tetracycline antibiotic. Table 1 documents which antibiotics were prescribed to these 960 patients. Using these 960 tetracycline patients as our “at-risk” population, we estimated the incidence of PTC-T to be 45 PTC-T patients ÷ 960 tetracycline patients ÷ 8 years of study x 100,000 = 586 per 100,000 person-years.

Table 1:

Tetracycline-Class Antibiotics Prescribed to Our 960 At-Risk Patients. We queried the Federated Utah Research and Translational Health e-Repository (FURTHeR) to determine which antibiotics were prescribed to University of Utah Health system patients ages 12–50 during the years 2007–2014.

Demeclocycline	Doxycycline	Minocycline	Tetracycline	Tigecycline
5	792	120	32	11

Sensitivity Analysis

The Moran Eye Center at the University of Utah is the only tertiary eye clinic in the state. For this reason, we are confident that most Utah patients with PTCS and PTC-T during the study period were seen at our institution. However, we only captured the tetracycline prescriptions that were written within the University system, and calculation of incidence is an overestimate. We were only able to obtain electronic health record data for the last two years of the study, 2013 and 2014. During these two years, the University of Utah health system provided care to approximately 11.05% of the Utah population (University of Utah Health Electronic Health Record; last accessed July 10, 2020). During these two years, U.S. Census Bureau data documented that the average population of Utah was 2,922,072. Assuming the population of Utah resembled the University system’s patient population in the proportion of those who were prescribed a tetracycline-class antibiotic, and in proportion of the population in the 12 to 50 years age range, we estimated the total number of patients at risk in Utah to be 8,809 (Table 2).

Table 2.

Determination of At-Risk Patients. Because we acknowledged that our estimate of the incidence of tetracycline-induced pseudotumor cerebri syndrome (PTC-T) was an overestimate, we estimated the number of at-risk patients in the state of Utah.

	Population served by University of Utah Health system	Utah Population
Total Population (all ages)	2,547,636 (8-year total) in the FURTHeR database	2,922,072 from the U.S. Census Bureau 2,922,072 × 8 study years = 23,376,576 (8-year total)
Population age 12–50	876,358 (8-year total) in the FURTHeR database (34.4% of patients in the database)	23,376,575 × 34.4% = 8,041,278 (estimated 8-year total)
Ages 12–50 years and prescribed a tetracycline	960 (8-year total) in the FURTHeR database (0.11% of the patients in the database ages 12–50 years)	8,041,278 × 0.11% = 8,809 (8-year total)
Ages 12–50 with PTCS	274 (8-year total)	274 (8-year total)
Ages 12–50 with PTC-T	45 (8-year total)	45 (8-year total)

FURTHeR = Federated Utah Research and Translational Health e-Repository; PTCS = pseudotumor cerebri syndrome. Percentages have been rounded for simplicity.

Using the estimate of the at-risk population for the state, we then re-calculated the incidence of PTC-T, the incidence of PTCS, the Risk Difference and the Risk Ratio (Table 3). This analysis estimated the incidence of PTC-T to be 63.9 per 100,000 patient-years and the incidence of PTCS to be 0.4 per 100,000 patient-years (Risk Difference = 0.0050539; 95% CI (0.0035728, 0.0065351); P < .001; Risk Ratio = 178; 95% CI = 130, 245).

Table 3.

Sensitivity Analysis. Using our estimate of at-risk patients for the state of Utah, we performed a sensitivity analysis to correct our overestimate of the incidence of tetracycline-induced pseudotumor cerebri syndrome (PTC-T).

	Prescribed Tetracycline
	Yes	No
PTCS (8 years)	45	274 – 45 = 229
At-Risk Population (8 years)	8,809	8,041,278 – 8,809 = 8,032,469
8-Year Incidence	45 / 8,809 = 0.00510841	229 / 8,032,469 = 0.00002851
Incidence Rate, per 100,000 person years	(45 / 8,809) ×100,000 = 510.8 per 100,000 person yrs*	(229 / 8,032,469) ×100,000 = 2.8 per 100,000 person yrs
Incidence Rate Ratio = 179 (130, 246), p < .001

FURTHeR = Federated Utah Research and Translational Health e-Repository; PTCS = pseudotumor cerebri syndrome. The incidence rate per 100,000 person-years is also an estimate of the one-year incidence.

DISCUSSION

The incidence of IIH among the general population is estimated to be between 0.65 and 0.9 per 100,000 person-years. ^{14, 15} We estimated the incidence of PTCS among patients taking tetracycline antibiotics to be at least 63.9 per 100,000 person-years, markedly greater than the estimated incidence of IIH in the general population.

Eldweik et al. (2019) recently reviewed claims data from a nationwide insurer and found that in an unadjusted, multivariable Cox regression model, there was a 70–91% increased risk of papilledema or PTCS in tetracycline antibiotic users.¹⁷ After adjusting for confounders, the increased risk was no longer statistically significant. The methodology used was quite different from that employed here, and claims data lack clinical data. PTC-T likely represents a spectrum of disease in susceptible individuals:¹³ There are some patients who develop PTC-T, discontinue the antibiotic and never develop PTCS again. There are other PTC-T patients who discontinue the antibiotic and improve, but later develop PTCS. This persepective on PTC-T likely explains, in part, the different conclusions presented by Eldweik et al. and the conclusions presented here.

Although a preponderance of our PTC-T patients had been prescribed minocycline, it is not possible for us to determine if this antibiotic is more likely to induce PTC-T compared to other antibiotics in this class. It is interesting however, that the tetracycline use pattern among PTC-T (37/45 = 82.2%) differs quite markedly from the prescription pattern in the FURTHeR query (120/960 = 12.5%). The authors suspect that this association with minocycline reflects its preferred use in the treatment of acne vulgaris,¹⁸ and not an increased propensity for this antibiotic to cause PTC-T compared to other tetracyclines.

A causative link between tetracyclines and PTCS would be best established with a randomized study, but for a number of reasons, it is unlikely such a study would ever be conducted. We conclude that the increased incidence of PTC-T compared to the incidence of IIH further strengthens the association between these antibiotics and PTCS. The reason for this association remains poorly understood, as does the pathophysiology of IIH. Some studies suggest a genetic predisposition to tetracycline-induced PTCS, with possible effects at the arachnoid villi, choroid plexus, or venous capillary bed.¹⁹ Retinol and retinol-binding protein have been implicated in the pathogenesis of IIH²⁰, and retinoic acid catabolism is inhibited by minocycline.²¹ Regardless of the underlying pathogenesis of the disease and the contribution by tetracycline antibiotics, our data suggest that the use of these medications puts patients at increased risk for the development of PTCS.

Limitations

In our sensitivity analysis, we considered limiting the number of PTC-T patients to 40, which is the number of those with Utah addresses. However, to do so would have required that we also limit the number of PTCS patients who did not receive a tetracycline to those with Utah addresses. By considering all 45 PTC-T patients, we have pushed the result closer to the null hypothesis of no difference. The sensitivity analysis also assumed that any patients in the FURTHeR query who developed PTC-T or PTCS were seen at the Moran Eye Center. The analysis also assumed that the proportion of Utah patients who were between ages 12 and 50 is the same as the proportion of patients of the same age who were in the FURTHeR query, and that the proportion of patients in Utah who received a tetracycline antibiotic is the same as the proportion of patients in the FURTHeR query who received a tetracycline antibiotic.

Based on our observations, we recommend that physicians who prescribe these antibiotics consider educating their patients about this adverse reaction. Patients who are taking these antibiotics and who experience symptoms of increased intracranial pressure (new headache, pulse-synchronous noises, transient visual obscurations, double vision) should be counseled to contact the prescribing physician for further advice. Patients who report these symptoms while taking these medications should promptly be evaluated by an ophthalmologist, who can determine if patients have ophthalmic manifestations of increased intracranial pressure.